Digital Transformation Review 11 – Artificial Intelligence Decoded
Artificial Intelligence – read this latest review whether you are a proactive adopter or watchful bystander!
Click to download the full report
The 11th edition of Capgemini’s flagship publication, the Digital Transformation Review, focuses on Artificial Intelligence (AI). It’s a topic of high interest to both consumers and enterprises alike.
This edition – Digital Transformation Review: Artificial Intelligence Decoded – presents a nuanced perspective on AI to help cut through the hype and fog.
We spoke to leaders and experts from a broad spectrum: large corporates, tech startups, academics, venture capitalists, and technology vendors. Often an entry point for organizations looking to explore the benefits of artificial intelligence, we also cover our research on voice assistants.
“AI will be the most debated, invested in, and disruptive business technology trend over the coming years. Organizations must cut through the fog and understand how to apply this innovation and become a truly intelligent enterprise.” Lanny Cohen, Chief Innovation Officer, Capgemini.
Explore the varied perspectives in the sections below:
AI: Augmented Intelligence matches people and machines
Michael is the Global Head of AI in the Group Digital team of Prudential plc. He joined Prudential from Silicon Valley based Pivotal Labs where he led the Data Science team. His experience lies in the application of artificial intelligence methods to large-scale, multi-structured data sets, in particular neural network based deep learning techniques. Michael previously founded and sold a ‘Silicon Roundabout’ based startup and prior to that was a partner at a major consulting firm.
Michael holds a PhD in theoretical physics from the University of Cambridge and is a Fellow of the Royal Statistical Society. Capgemini’s Digital Transformation Institute spoke to Michael to understand how AI is helping Prudential in a variety of ways already.
AI at Prudential – much more than a cost play
What does AI mean for you?
In her book Artificial Intelligence, Elaine Rich says that AI is the study of how to make computers do things that people, for the moment, are better at.[1] This is the number one thing for me: it is not about automation. When people talk about automation, they largely do so with cost in mind. AI is more importantly about scalability and customer experience primarily, and given we are in financial services, about regulation and compliance. We want to be compliant by design, and that is much more easily achieved when you have computers do tasks than humans. AI is all about customer experience, scalability, compliance by design, and then cost. In this order.
What are some of the AI Initiatives that you have launched?
We have developed a robo advisor, which we launched in our Taiwanese business early this year. As an organization, we employ around 600,000 financial advisers around the world. We already have a lot of information and historic data about the customer and financial markets. We want to use this data to provide sensible and tailored suggestions to our customers. It’s important to point out that we don’t want to replace our advisors – they are a really important route to market for us. But we want to augment them and add to their capabilities. We want our human advisors to become prime users of our robo advisor, so they can tune in even better to the actual needs of their customers. What we really want to do is to use humans to the best of their capabilities. AI is taking away the time humans previously spent on repetitive issues and allowing them to focus on where human intelligence can drive value – for both themselves and for customers. It is also why we call AI ‘augmented intelligence’.
Employees welcome AI that helps them
How have employees reacted to these technologies?
We are clear on what we want to achieve – we have no interest in killing jobs. We know that the people who work for us are extremely well-trained, understand our products, and understand our customers. We want to use those team members to the best of their capabilities. When we implemented an intelligent voice box for compliance purposes, we expected a lot of apprehension and resistance. Surprisingly, employees were quite enthusiastic. The reality of their job is that they spent a lot of time on real humdrum issues that really don’t require any of their skill and knowledge. For instance, they spent four minutes on every call just doing identification and verification of the caller. And they told us – ‘look, if your box gives us those four minutes of a call back, that would already be an amazing advantage’.
Selling AI into the business: prototyping and a focus on demonstrable successes
Can you give us a sense of the investment that was required before you saw benefits and returns?
‘No’ is the simple answer, though not for reasons of confidentiality. It is because we do not have an overall business case for artificial intelligence. We try and build ourselves into every single project and activity that is going on and influence the business case as it happens. In a highly federated organization such as Prudential, our role is to ensure that the business owners see the value in what we are trying to achieve and fund it.
How do you convince business owners of the merits of investing in an AI-driven solution?
We have a three-step process. Firstly, we have a strict no PowerPoint policy. If we are talking to an executive, we need to have a working demo. Secondly, we build on existing successes that we’ve had in other parts of the organization. And the third, and most important part, is how we make rapid prototyping a key part of our approach. We put together a “hot house” where we pull a team of people from across the organization, give them three days and a set task and have daily demos. The team comes up with a draft working prototype that the judges of the hot house – typically the business owners of the problem – can see. They can then take a clear call on whether it works or doesn’t. If it is on the right lines, we get the remainder of the 90-day cycle to develop it and make sure it actually works. This is exactly how we got the go-ahead for our regulatory voice box initiative.
Biggest challenge – getting people comfortable with robots
What were the biggest challenges you faced in rolling out these initiatives?
There are two major challenges.
Firstly, how to deal with robots? How to get used to the real-time nature of using software and seeing feedback coming in after changing a few parameters? How to overcome skepticism towards AI-driven solutions?
The second challenge is more technical in nature, e.g. take a call on data availability, or the platform/ technique you want to use.
Do you see the need to develop talent and skills as a key challenge?
We have a structured approach to enhancing digital skills. For instance, we are running a training program for employees from all BUs to learn Alexa programming skills. The primary objective is not to develop AI solutions, but we are trying to increase the level of confidence that our colleagues have with AI. We hope to build an understanding of what AI both can and cannot do, as they are equally important. But one thing we are not going to do is go on a big hiring spree and try to hire up all the AI experts around the world. Instead, we will focus on hiring people with the right mathematical background and aptitude to understand our problems, our data, and our customers.
[1]Artificial Intelligence, Elaine Rich and Kevin Knight, 1991.
Atif Rafiq is the Global Chief Information Officer and Chief Digital Officer of the Volvo Car Group. At Volvo, Atif is spearheading the digital transformation of Volvo Cars by building new consumer experiences and the digital products behind them. Atif joined Volvo Cars from McDonald’s where he was a Senior Vice President and the first Chief Digital Officer with a focus on the future of customer experience and brand engagement, powered by digital technology. Earlier, Atif was the General Manager of the fast-growing Kindle Direct Publishing unit at Amazon helping give shape to the future of content and books.
Atif also leads Yahoo’s Local Business and managed Product Marketing/
Strategy for Yahoo News. Atif began his long career of almost two decades in Internet businesses starting with AOL and Audible. In 2001, Atif co-founded and served as CEO of Covigna, an innovative content management software provider until it was sold to Proquest. Atif holds a university degree in Mathematics Economics and an MBA in Finance and Marketing.
Capgemini’s Digital Transformation Institute spoke to Atif to understand his views on Artificial Intelligence and how it impacts the automotive industry and Volvo Cars.
How does Volvo Cars see artificial intelligence (AI)? Are you already deploying it, and if so, in what functions?
Car companies are actively using AI in their autonomous driving efforts to solve very complex problems around being able to take you from A to B in a safe way. This typically gets the most headlines. And it’s been used for a long time in driver assistance and active safety features of the car, like for instance traffic sign recognition and threat assessment. But every facet of this industry can benefit including how cars are made and sold and to invent new
customer experiences.
Let’s take the process of trading in a car and the need to describe the condition of the vehicle. Objectively determining the condition can be a machine-learned activity that makes things easier for everyone and perhaps increases trust. This can work by taking images of your car’s conditions and matching them against large data sets of cars categorized according to different physical conditions. The same goes for returning a rental car accessed through a sharing service. For example, determining the car comes back in the same condition compared to when it was checked out.
Another useful application, currently under development, is predictive maintenance. Parts within your vehicle are designed to work well for a long time. Now, with AI we can observe signals about how they perform against data sets of repairs to predict a potential need for maintenance before anything stops working properly. This clearly holds
huge benefits for the consumer, where we can optimize the service intervals of the car and avoid any unnecessary downtime. All of that leads us toward the goal of a hassle-free ownership and better experience for the driver.
As you can see, our aim is to use AI to solve meaningful customer and enterprise problems anywhere it makes sense.
Where do you think AI can help a large auto manufacturer such as Volvo Cars?
We are working on uses for the consumer and for the internal enterprise. Both opportunities are significant. Let’s start with our vehicles. To make good use of your time in the car, you might want to operate different features with your voice. We’re making strides here by being one of the first companies to partner with Google on a native version of Android in the car that makes it easy to use Google Assistant to access all kinds of information and services. We
realize there are digital assistants like Amazon Alexa and we will find the best ways to support them.
Then of course with the car we have self-driving both for consumers who buy our cars and B2B customers like Uber which seek to operate fleets of autonomous cars. Both of these are major market opportunities for us.
And we can make an endless series of experiences around owning or using a car easier. My favorite at the moment is the idea of using exterior cameras of cars to figure out when your car needs a wash. This is definitely doable though hard to get right because, as you can imagine, everyone has a different perception of cleanliness. Wouldn’t it be great to learn how each person wants the appearance of their car to be and to simplify life by dispatching a cleaning service to keep it that way?
The good news is we have the building blocks here with our on-demand wash service. Customers in certain cities can use the Volvo on Call app to request such a service. These kinds of ideas are not technology for technology sake. They can help us fulfill our promise to give one week of quality time a year back to consumers in 2025.
Since I also run global IT at the company, I’m happy we are putting effort into enterprise use cases. Those are just as plentiful when we think about making cars, the amount of automation in manufacturing, production, and logistics.
What is your view on the skills issue with technologies such as AI? Are you finding good AI talent easily?
The skills involved with AI and machine learning are in demand, yet we have been successful in recruiting talent. That’s because our industry is one of the best places for real applications that will influence how millions of people live. There is so much technology in cars—sensors, cameras, radar, etc.—generating so much data with increased compute capacity.
People see this as a rich environment in which to work. There’s a unique opportunity right now to solve fundamental technology problems. Fortunately, we’ve been able to recruit talent in both Sweden and Silicon Valley (through our digital tech center).
Of course, we can do better. Part of the approach needs to be sharing ideas and socializing the work the company is doing so we get noticed. I think we will do more of that as we look forward through conferences and events.
How is AI being driven in Volvo? Is there a central leader for all initiatives?
There are some dedicated efforts as you can imagine like autonomous drive that we resource through our JV with Autoliv in a company called Zenuity. Zenuity is a separate company building autonomous drive software and they employ a large number of software engineers, including those familiar with AI.
More broadly we have a Center of Excellence (CoE) that brings together different skills from algorithm development to data science to setting up compute environments. We formed this CoE over the past 12 months and this group shares learnings, applies its efforts to different parts of the company and
will find ways for us to place AI in the heart of our development activities everywhere in Volvo.
A good objective for our company is to have AI models for all software development domains where it applies, just like Google has a clear method to “AI-ify” every product from mail to photos to maps. And to do that we will need to drive into the core development processes. This is a big ambition and I’m hopeful we will take solid steps in the next 1-2 years.
What is your take on the debate of the impact of AI on jobs?
There are a number of offsetting factors when it comes to the use of AI. The question is how will it net out?
If AI eats jobs at a rapid rate without producing new work for managers in corporations, that will make companies more profitable but hurt society. Tech companies are keenly aware of this dynamic and so we see conversations about universal basic income. I think Amazon’s move into healthcare, described as a not for profit motivation, can be a positive offsetting factor to potential job loss.
My own take is that we will see AI live alongside humans in carrying out managerial work and that AI will assist humans to unlock human potential, allowing them to focus on higher and higher value tasks and activities. The work of humans in corporations is going to be very hard to replace for a long, long time but it will need to rise to higher and higher levels of contribution. AI can help make the workplace more objective in decision-making based on real organizational learning.
Volvo has publicly said you want to put driverless cars on the road by 2021. You’ve also partnered with Uber. What role does AI play in driverless cars?
AI is the brain that takes advantage of the car’s ability to sense the world through sensors, cameras and other hardware like LIDAR* to make decisions. These decisions must be accurate, fast, and continuously improved so that we make driving safer than it is today. Of course, AI is at the heart of that.
Volvo has historically been known as a brand that stands for safety. How is AI helping you there?
Since the transition of focus from passive to active safety, information technology has been at the core of what makes up safety development. Today, AI assists in this development in two fundamental ways: solutions to understand when a crash is imminent to trigger pre-emptive safety systems and, second, increasingly to avoid accidents by understanding the environment and anticipating problems before they happen.
Within safety, AI can also considerably enhance the tools and methods used during the development phase. One example is helping us to understand complex situations with a lot of interacting agents, which is typically the norm when it comes to this area of engineering. I think we will see more simulation software based on AI that allows for modeling complex scenarios and some unsupervised learning insights.
You have expressed a vision where technology dramatically simplifies life for the car owner. Can you describe it and what if any role AI plays in actualizing this vision?
We have a publicly stated goal of giving our consumers one week of quality time back each year by 2025. That can come in many ways.
For example, by creating a car experience that works well with the rest of your digital ecosystem. Cars can do much better here and that’s why working with a company like Google on native integration of Android will connect things in very interesting ways. Knowing your calendar and where you have to go can be used to save people time in all sorts of ways. For example, imagine the case where you have an early morning meeting, the fuel level in your car is low and the distance to get to the morning meeting is a bit far. Why should this stress you out in the morning?
There’s enough knowledge between the car and your calendar to solve this problem the evening before your meeting and reduce morning headaches. We talked about dispatching fuel on-demand and this one case where you might appreciate it.
Speaking of dispatching fuel, we are testing a number of convenience services like valet pick-up of your car for service or repair, test-drives at home, and even in-car delivery of packages (sometimes that is safer or more convenient). For Volvo and its partners to provide these services, we have to intelligently route resources like people and cars around. You can imagine AI providing a smarter basis for these services that improves utilization and efficiency of the resources.
Coming back to the customer, owning a car used to be about freedom. Technology is going to bring that freedom back by removing headaches of managing tasks like these, where AI can be an enabler behind the scenes.
You have an uncommon role for the automotive industry—a joint CIO/CDO. What advantages do you think that brings to accelerating digital initiatives?
Typically, a CDO focuses on consumer facing innovation and technology, while a CIO oversees the digitization of the internal enterprise. You can imagine there’s a good amount of dependency between the two, such as platforms that serve both consumer and enterprise use cases.
One example is the connected car platform, which is used by consumers to handle remote tasks like heat, locking, and sharing, but also for internal use like collecting information about the performance of the car for maintenance. Another example is data platforms which house customer information used in retail interactions for sales and service, typically powered by enterprise systems. This platform is also used when consumers interact with us directly over the web or mobile. By combining the consumer and enterprise digitization efforts in my role, we aim to leverage common capabilities more effectively.
On top of this, we think about digitization as a fundamental change in how work gets done through real agile processes. To do this only for consumer technology and software and not for enterprise is short sighted. We are equally excited by the potential for digitization to fundamentally change how software gets developed, managed, and
driven by our strategy both for the consumer and the enterprise.
Lastly, I think this combination of responsibilities make sense for all kinds of industries because customers don’t care how your company is organized. It’s odd to settle for one part of a company moving at a slower pace than other parts; the ambition must be bigger across the board.
Professor Luciano Floridi is the Oxford Internet Institute’s (OII) Professor of Philosophy and Ethics of Information at the University of Oxford, where he is also the Director of the Digital Ethics Lab. Outside Oxford, he is Faculty Fellow of the Alan Turing Institute and Chair of its Data Ethics Group; and Adjunct Professor of the Department of Economics, American University, Washington D.C.
His research concerns primarily Information and Computer Ethics (aka Digital Ethics), the Philosophy of Information, and the Philosophy of Technology. Other research interests include Epistemology, Philosophy of Logic, and Philosophy of Science.
Capgemini’s Digital Transformation Institute spoke to Professor Floridi to understand his views on ethics in AI and the issues that organizations need to bear in mind when designing AI solutions.
Pace of change is the bigger challenge
AI is most often equated with job losses. What is your take on this debate?
I think there are a distraction and a real issue here.
The distraction is the fear that jobs will simply disappear completely and that this will be a major disaster. Why do I think it is a distraction? Because it’s based on two fallacies.
The first fallacy is that there is only a certain fixed amount of work to be done. This is not true. There is as much work as you want to do, depending on resources, time, who does what, skills, and so on. The example I have in mind is how much work you can do to clean your house. It’s bottomless. It’s just that at some point you draw a line and decide that it’s clean enough. So, there is no fixed amount of work as such.
The second fallacy is that AI will simply replace work, when it will also make work surface that was not economically viable to do in the past. For example, if I buy a new robot to cut the grass, the gardener will finally have time for the roses. Earlier, having time for the roses was not viable because it was too expensive. Having the robot cut the grass makes the job of taking care of the roses economically viable.
If you consider these two issues – fixed amount and viability – then you know that AI impact is actually much more complicated. The oversimplification is the distraction.
So, what is the real problem then?
The real problem is not the replacement in itself. It’s the pace at which the replacement is happening now and will continue to happen. Technologies in the past have replaced jobs slightly more slowly than anything we are seeing today. Today, from literally one year to the next, some jobs will become totally obsolete. If that is the case, how you re-train and re-skill the workforce – and how you develop social support systems to mitigate the impact – are the real issues. Future generations will have new jobs. There is no doubt about this. Computers will require human beings to handle them. Look at Audi for example. They have a one-to-one ratio – every robot introduced requires a human being. So, more robots may actually require more human workers. I am not worried about future generations, but more about us – the generation that is undergoing the shift. This will be traumatic, but more because of the pace rather than the nature of the phenomenon. Society needs to help those who will feel the brunt of the drastic changes.
What kind of timeline do you believe we are looking at when it comes to large-scale impact on jobs?
We will see the impact in the next 10 to 20 years, but exactly when depends on many variables that are highly unpredictable. For example, will we see an AI backlash – an opposition to AI similar to what we saw against genetically modified crops? The potential social reaction and legal impact are still very unclear, and this could affect the impact of AI quite dramatically. Will we see regulation, for example, on when automation is allowed in certain contexts? Think of all the current legislation we have when it comes to security in public transport. For example, the law may still require to have drivers on board of busses and taxis, as it does for airplanes. We might see similar legal frameworks with AI and automation.
But I would say that, within 20 years, the world will have profoundly changed.
You mentioned re-skilling. Is it up to companies to re-skill their staff?
This is a very important question. If we talk about skills that you acquire at an early stage of your career, this is very much a joint effort between companies and broader society, including the educational system. But when it comes to re-skilling someone in their fifties, for example, it seems to me that this is more on the societal side and less on companies. It is up to our society to help with the transition.
Companies should think of AI as a reservoir of smart solutions
In terms of organizational impact, will AI truly change the structure of large companies?
In the next three to five years – at the point when AI gets into the pockets of ordinary citizens – I suppose companies will start to experience AI on tap. The analogy here would be with cloud computing or electricity. You don’t produce your own electricity, you just take advantage of it. Likewise, you might just take advantage of smart solutions that can be deployed to solve specific problems. Companies that start thinking in terms of AI as a reservoir of smart solutions are going to be better placed than others to take full advantage of the new digital transformations. Even the structure of organizations will be affected by where they can deploy these AI solutions.
Which industries are best placed to benefit from AI?
The medical sector is in for major changes due to AI. I also believe AI will impact the security and safety sectors significantly. These include anything that needs supervision 24/7, such as monitoring or prediction of possible faults in airplane engines, or early signaling of potential threats. Basically, anything to do with management, safety, prediction and optimization becomes more efficient with AI.
Adapt AI to humans, not the other way around
What are some of the ethical considerations that organizations need to consider as they implement AI systems?
I fully subscribe to the usual discourse around privacy and protection of personal data with AI. At the same time, I believe that is not complete.
My first concern is that once we have real, everyday AI, we need to make sure that the design of smart environments does not result in us always being the ones who adapt to AI rather than vice versa. We are currently deploying smart agents that are rather rigid in what they doing. Think of a world where there are lots of Stage III or even Stage IV driverless cars. This will mean we will see humans adapting to artificial agents. While this is probably inevitable, I would say it is an ethical imperative to make sure that the malleable, adaptable and intelligent human agent in the partnership is not the one that adapts all the time, compared to the stubborn, hard-working and rigid AI agent.
Secondly, we are surrounded by agents that are gently nudging us in certain ways – to take our holidays abroad or read the next Harry Potter book. This constant gentle pushing and pulling is definitely affecting us. These constant reminders and suggestions are shaping us. How do we make sure that we are aware of that? In other words, that we know what we are doing and that this gentle nudging is reduced? We need to protect our autonomy.
What role should governments play in developing policies for these checks and balances?
I think the checks and balances are essential. I would like to see a normative framework with maybe an ombudsman or organizational self-regulation for AI. We need an authority where we can go and say ‘there was a mistake’, or ‘it wasn’t me’, or ‘can we check this because this isn’t quite right?’. Prevention and redressing of problems caused by AI must go hand in hand.
What AI application do you think will transform people’s lives?
If you look at good consumer technologies, they fall in two camps.
First, there are things that we do, but don’t want to do. For instance, imagine if AI can help us with a robotic arm that puts the dishes inside the dishwasher and take the dishes out of the dishwasher and back into the cabinet. That is a category of application that’s useful. So, anything that does things that we are forced to do these days, but we don’t have to do anymore, would be a great success.
Second, there is a category of things that we wish we could do, but we cannot or didn’t even know. Once a technology solved the issue, it dawned on us. A smartphone is not something that takes away something we do not want to do, it gives us something more to look forward to. In both cases, a good way of assessing an AI application is not whether people initially buy it, but whether they buy it again when it breaks down. So, the AI that will be successful is not the one I want, it is the one that I want again because of what it makes me not do, or what it enables me to do.
Michael Schrage is a Research Fellow at the MIT Sloan School’s Initiative on the Digital Economy. He is the author of the books Serious Play (HBR Press), Who Do You Want Your Customers to Become? (HBR Press) and The Innovator’s Hypothesis (MIT Press). His latest book, The Innovator’s Hypothesis, looks at how to develop innovation in a quick, cost-effective way. Michael is a columnist for Harvard Business Review, Fortune, CIO Magazine and MIT’s Technology Review, and is widely published in the business press.
Capgemini’s Digital Transformation Institute spoke to Michael to gather his views on what AI means for large organizations.
Machine Learning more important than traditional AI
How would you define Artificial Intelligence?
We first have to distinguish traditional AI from machine learning. Many people understand “Artificial Intelligence” as software that appears to replicate, mimic, or effectively copy the human thinking or cognitive process. With ML, the algorithm is capable of learning both in a supervised and unsupervised way, based on the way it is trained on data. I believe that machine learning is really becoming a much more important pillar and organizing principle in algorithmic, systems, and process design. We see this in companies like Google, Facebook, and Netflix. We see this emerging in organizations that are exploring the Internet of Things.
Where do you think we are with the development of AI?
What AI already allows most organizations to do today is hollow out or replace ordinary cognitive work. On average, it is more economical, and makes more business sense, to invest in AI system that reliably manage data and processes in ‘above average’ ways rather than invest in ‘typical’ human beings or a groups. I believe that at the very top of the pyramid there are still measurable and significant human-being advantages. But right now, with AI in combination with machine learning programs, you can design business processes where the system consistently, reliably, and cost effectively outperforms its human counterparts. We see this in financial services; we’re seeing to more in manufacturing and industrial controls; we’re certainly seeing it in IT/digital services. It is a very bad time to be average in performance. Even being above average no longer commands a market premium.
AI offers significant opportunity for data leaders in a compliance-oriented world
Where do you think AI will have the most impact?
AI will be quicker to enter industries that are heavily regulated. That’s because compliance and regulatory-driven industries are very complicated, but they generate lots of data. Also, compliance costs is an overhead cost that must be reduced. So, wherever there is intense regulatory oversight – such as finance, energy, pharmaceuticals, etc. – you will see disproportionate investments in AI and ML.
How can organizations realize the transformative impact of AI?
Companies need to manage data as an asset to get meaningful and measurable economic returns. Start there. Every company is made up of three categories of data managers. Immature managers ask, ‘how do we manage this data?’ More mature managers ask, ‘how do we get value from this data?’ And superlative managers say, ‘what kind of partnerships, governance, and technological investments do I need to make to get best-in-class returns on this asset?’
AI success is a question of leadership
What separates companies that make a success of AI from the laggards?
AI leaders have a policy and process around data governance and treat data as an asset. How they oversee, and share data is critical to both efficiency and growth. They also have key problems or business cases that lend themselves to known structures for AI and ML algorithms, such as intrusion detection systems, trading in financial services, predicting churn from subscribers or identifying certain kinds of customers. These companies view AI as an enabler and they are ready to experiment. They’ve got the digital/IT infrastructures to do so.
How should organizations run AI initiatives – top-down or bottom-up?
Pockets of excellence almost always exist so that’s a tough question to answer without knowing the firm. Many financial services companies, for example, are doing excellent machine learning work and data science work. But it becomes like a ghetto. They cannot move it to other parts of the organization, because the other parts of the organization don’t understand it, or don’t have the talent, or are concerned that people might lose their jobs. These are human issues that have nothing to do with the capabilities of the technology and everything to do with the culture of the organization and the quality of its leadership.
The future of job and work
How do organizations get the right AI talent, and can they upskill their existing people?
Are organizations able to hire data scientists versus organizations like a Google or Netflix? If you are a tier-1 company, you probably have a decent chance at getting quality data science, machine learning and AI folks. But if you’re a tier-2 company, what do you offer besides lots and lots of money?
Organizations need to know what it costs to turn a good 40-year-old coder and developer into an ML or AI coder and developer. What portion of mid-career people can be converted in such a manner? I have no good answer to that. But I think that is the dominant human capital question going forward. I worry less about folks in their 20s, I worry about people in their 40s.
What kind of jobs do you think are at immediate risk of being disrupted by AI?
At this stage it is still not clear whether the level of job generation will match job destruction. I don’t know. It is challenging to say whether multiple professions will go away or whether how you create value in those professions fundamentally changes. The name may be the same but what you do and the technology you’re doing it with are different. For instance, the idea that radiologists in the year 2025 will be doing largely what they did in 2005 strikes me as difficult to believe. Similarly, in America, it is easy to imagine scenarios where judgments for certain kinds of arbitration or civil lawsuits are rendered algorithmically rather than by human beings making arguments in front of a judge.
More specifically, I believe that the compliance staff at pharma and financial services companies will be cut by between half and two-thirds within five to seven years. This is precisely because large organizations are using AI and ML to manage the compliance process and the lawyers who remain will largely supervise algorithmic outcomes rather than people. You won’t need nearly as many people do it.
Battle between process efficiency and user experience
How do you see AI evolving in the future?
I see a big battle in ML and AI between process efficiencies and user experience. Amazon, Netflix and Google are good examples of organizations that try to balance user experience with process efficiencies. The financial services companies and industrial companies are torn, because their first goal is improving their return on assets. This brings in a bias in their usage of ML and AI – how do we improve our internal process efficiencies as opposed to giving our customers and users a better experience and better value? I think that will be the great schism in the market over the next 5 to 10 years.
Frank Chen is a partner at the venture capital firm Andreessen Horowitz. He runs the research and deal team at the firm, which systematically identifies and evaluates investment opportunities, and builds knowledge about technology, people and trends.
Prior to Andreessen Horowitz, Frank was the vice president of Strategy for HP Software, where he helped the company understand and act on changes resulting from the rapid enterprise adoption of virtualization technologies across servers, network and storage. Frank joined HP Software through its acquisition of Opsware, where he was the vice president of Products and User Experience for a broad set of Opsware’s data center automation products.
Frank holds a B.S. in Symbolic Systems from Stanford University, where he graduated with distinction and was elected to Phi Beta Kappa. Capgemini’s Digital Transformation Institute spoke to Frank to understand how he sees AI as a partner at one of Silicon Valley’s most respected VC firms working with some of the hottest startups.
AI is here, and it will impact everyone
AI has been around for a while. Why is it different now?
AI is now inside a bunch of mainstream applications that everybody uses from Siri to Pinterest. This is a relatively recent phenomenon. Probably a quarter of applications have AI today and it’s growing fast against maybe less than one percent 20 years ago. The big change from a technology point of view has been the ‘deep learning’ machine learning algorithms in conjunction with the availability of lots of data and the ability to crunch that data on many different computers. That is what has made the difference since when we started in 1956.
Where will AI be most useful?
Asking where AI will be most useful is like asking which sectors were going to benefit when database technologies emerged. It turned out the answer was every single one of them. It’s hard to imagine a piece of software where you don’t have a database behind it. They were good for United Airlines, good for Starbucks, and good for Exxon. AI is exactly like that. Companies need to be thinking about what to build, how to price, how to reach customers, and how to support them. All of that is going to be turbo charged by AI.
What are some of the ways in which AI can help organizations?
The way to think about AI is it can do lots of things that previously you would give to humans to do. Imagine that on every customer support call you could have someone listen in and they could tell you what product or services the customer is interested in or whether the customer is getting angry. You could hire an additional rep to listen in on every call, but it wouldn’t be cost effective and nobody does it. They end up recording all their calls and then they don’t listen to those calls. So, AI can listen to those calls and answer all of those questions for you: are they getting angry, what products should we recommend to them at this point, are they about to turn and if they’re about to turn what can we offer them, so they don’t? Think of it as having an army of interns that can listen in to calls, read documents, look at pictures, or make predictions based on past history. That’s what AI is going to be able to do.
Are you seeing a growing number of startups offering AI solutions?
Yes, basically all the startups that we find these days are AI startups. When we first started as a firm in 2009 nobody called themselves AI-powered. And flash forward to today, 80% of the startups we see call themselves AI-powered. So, it just shows the sea change that’s happening. AI is rapidly becoming the must-have technology – like the iPhone app seven years ago. When we see a startup that doesn’t have machine learning and AI, the first question we ask is ‘so who’s doing it with AI?’
The talent gap is only temporary
When the top tech firms are scooping up all the hot AI talent, what should traditional organizations do?
Every time there’s a platform shift, there’s a temporary shortage of talent as the ecosystem catches up to provide the top engineers with the right set of skills. When SQL first came out, there was a run on SQL developers. The universities weren’t graduating them out fast enough, and the first ones were super expensive. And the same thing was true when the iPhone came out, because Apple in those days required that you write all iPhone applications in this weird language called Objective C that nobody knew. The web was like that too – nobody knew HTML and JavaScript when the web first came out. And so, every time there’s a platform shift there’s this temporary period where the developers are expensive and where there’s not enough of them. But eventually the ecosystem catches up and it creates enough of them. So now there are tons of boot camps where you can go for three months and train yourself on the latest AI machine learning techniques.
When I took the CS221 ‘Intro to AI’ at Stanford in the late Stone Age, there were 50 people in the course. This semester there are three ‘intro to AI’ classes. Each one of them has 1000 students. For context, Stanford has something like 6000 undergrads total, so one-sixth of the entire undergraduate population is taking one of the three ‘intro to AI’ classes. And so, the ecosystem is catching up fast and we won’t have this sort of temporary imbalance of talent that we do now.
So, do you believe a combination of re-skilling/ up-skilling can help when it comes to developing AI talent?
Yes, that’s exactly right. Every university is adding this set of techniques to their computer science curriculum and there are dozens of boot camps. It’s also important to bear in mind that large traditional organizations don’t necessarily need the multi-million-dollar top AI researcher. They don’t need to invent a brand-new AI algorithm or technique, they can just take the ones that are being invented and apply them into their business. So, the great news is that those research engineers that go into Google, Apple, Facebook, Amazon and Apple and so on, they are open sourcing what they’re doing.
What I don’t want is an industry executive to read a news article and conclude that since they can’t afford a $10 million artificial intelligence researcher, they are just not going to do anything. That’s the absolute wrong conclusion to draw. The right conclusion is, ‘hey, those top multi-million-dollar researchers are open sourcing their stuff – that’s fantastic. I can take an image of it without paying a million dollars.
Sending motivated, curious people to get the Udacity Artificial Intelligence nanodegree is a 6 month, $1,600 investment. It’s not 2 years and $100,000.
AI-native organizations treat data very seriously
So, what are some skills that traditional organization should focus on developing?
Andrew Ng, who is one of the most well-known researchers in this space, has this really good analogy. In the early days of the Internet, even your neighborhood shopping mall had a website. But just because they had a website, that didn’t make them an internet company. The Internet retailer was Amazon. And becoming an Internet retailer required Amazon to master a lot of things. They mastered continuous development and deployment. They changed their website 100 times a day as opposed to the mall which changed it like once a year. They mastered A/B testing. So, they mastered all these things and that’s what made them an Internet native.
The same thing is happening with AI. Just because you use AI techniques doesn’t make you an AI native. As an AI native you’re going to have to master a whole different set of skills. We don’t know exactly what set of skills these are going to be yet, but there are a few early signs. An AI-native company is going to treat data very seriously – harvesting the data, feeding the data into algorithms, and labeling the data so that the algorithms can make accurate predictions.
Make AI a daily habit for your organization
What are some areas where partnership with AI-driven startups makes sense for large organizations?
Building AI tools requires consistency. You have to make AI a daily habit, because if you don’t make it a daily habit your competitor will make it a daily habit. There are startups that offer tools for people who are building AI applications.
If you have a model that makes predictions, SigOpt, for example, can make that model better by automatically tuning it. Every model has a dozen or more knobs that you can turn and, depending on where the knobs are set, the model produces different predictions. There’s an optimal setting for each knob. So, you can either pay a data scientist to basically trial the settings, or you can just use SigOpt, which will automatically put the knobs in the right place to make the best possible recommendations. So that’s a tool that organizations can use to get the expertise in-house.
What are some low-hanging use cases for AI?
The low-hanging fruit is basically anywhere where you’re making a recommendation, anywhere where you need to understand what people are saying, or anywhere where you need to recognize what’s happening inside a picture. So, for example, having a chat interface to your brand so that people can chat with you on an automated basis. Low-hanging fruit would also be understanding text, pictures, a video, and making recommendations.
What about the more complex use cases of AI?
We have a lot of biology investments that are basically doing such as taking blood samples inside your blood sample. In other words, free-floating DNA that’s not bound to tissue and is just floating around in your blood stream. We can take that DNA and use it to predict whether you are going to get a specific type of cancer. So, compare that to the global standard today, which is you need a tissue biopsy. There’s a company called Freenome that can do that without any tissue. For cancer, the most important thing that you can do is early diagnosis as survival rates are 90% when you find it early, but 10% when you find it late. So, the single most important thing you can do with cancer therapy right now is not to invent something for the 10% case. In general, most AIs are now likely to be as effective or more effective than the very best human diagnostics.
AI eats jobs, but fears of an AI takeover are overblown
AI is almost always spoken of in context of job losses. What are your thoughts on this?
Every category of automation eliminates jobs and creates jobs and this has been true since the very dawn of automation. When we invented the looms, a lot of weavers lost their jobs. Today, less than five percent of the population is involved with farming as compared to 80% at the turn of the century. My take on it is automation has displaced people, but it’s always created new opportunities and AI is just one in a long line of automation technologies.
We didn’t know we wanted iPhones until Steve Jobs showed them to us. Our capacity to want new things and new services is basically infinite. Therefore, if it’s infinite, then somebody is going to emerge and service that need and create new jobs around it. I don’t know why AI would be any different than any other automation technology. Some of the jobs AI will replace are also not work that anyone wants to do. For example, AI is going to eventually make it possible to automate strawberry harvesting. If you’ve been a strawberry harvester, you know that it’s a miserable job. In general, that’s good, as long as we can find the strawberry harvester another job, which I’m pretty confident that we’ll be able to do, because we always want new jobs and new services.
Premature to worry about super AI
Elon Musk worries about World War 3 when it comes to AI. Machine learning researchers disagree. What is your take on the debate?
To frame that answer, let me describe maturity levels for AI. Everything that I’ve been talking about is what the research community calls narrow AI, which is very specific. Pinterest built something that can look inside a pin and figure out what’s in it and what’s similar to it. Those algorithms don’t help Lyft to figure out the best route to take you from place A to point B. And then those Lyft algorithms don’t help Freenome figure out if you have cancer. They’re very task specific and that’s why they’re called narrow, they don’t generalize. Even in these narrow fields there are lots of things that people can do that algorithms can’t do.
The second sort of AI would be general AI. In other words, you and I can learn new things, like you can learn to ice skate, you can learn to do double-entry book-keeping, and you can learn to do machine learning and a lot more. AIs can’t do that yet and there’s no consensus in the research community about how you would build a machine that has this remarkable property that our brains have, which is the ability to do new things in new domains very quickly. So, we’re not even on a path to really do that.
And then the third stage is what Elon Musk is most worried about, which is super AI and ‘singularity’, where AIs are better than humans at everything. At that point, there’s no looking back, because computers are faster, have more memory, and are more repeatable, therefore making humans irrelevant. That seems very far away. As Andrew Ng says, it’s like worrying about overpopulation on Mars. One day we might have to worry about that, but nobody is on Mars yet, so let’s not worry about that now. Even in narrow AI there’s a long way to go, so worrying about super AI seems a little premature.
Rajen Sheth is Senior Director, Product Management currently responsible for running the Google Cloud Artificial Intelligence and Machine Learning product line. This includes products which bring Google’s AI technology to enterprise customers and developers, including Google’s AI platform, image, video, and speech AI capabilities.
Previously, Rajen was responsible for the Android and Chrome for Business and Education group, leading a cross functional team across product, engineering, sales, support, ops, and partnerships. Earlier, Rajen also co-founded and led the Google Apps product line, and Google’s messaging and collaboration products for businesses. He was also responsible for leading Google’s efforts to bring cloud computing to universities and enterprises.
Rajen holds a B.S. and M.S. in computer science from Stanford University. Capgemini’s Digital Transformation Institute spoke to Rajen to understand how Google is democratizing AI.
Can you tell us a bit about your role at Google?
I’m the Director for our Google Cloud artificial intelligence team, where we bring Google’s many AI technologies to bear for our customers and for developers. We do a number of things. At the infrastructure level, we bring the best of infrastructure for artificial intelligence to customers. At the developer level, we give developers access to a lot of the great models that we’ve developed for products as well as those built specifically for enterprise use cases. Thirdly, we provide solutions that meet specific business problems.
AI is going to be in every industry in ten years
What are some of the implications of AI for large companies?
I think the industry is still figuring out where AI can be most beneficial for companies. But I do see a number of areas. For example, we are seeing a particular interest and impact in personalizing customer interaction. It is always hard to give a customer a personalized experience when you have multiple customers that are coming to your website. AI can help give a much more personalized interaction -everything from recommendations to helping interact with the customer via customer service applications such as chatbots. AI can help in contextual impact, so that the customer can get the answer they need quickly. AI helps in marketing campaigns and loyalty programs as well. Another area is efficiency and agility. Spotting new patterns in a database can lead to strong efficiencies. Another area is adding structure to unstructured data.
Which sector do you believe will be most affected in the short term?
We are definitely seeing retail as an area where there is a lot of potential for AI. Everything from how you personalize a customer interaction all the way through to how do rethink the concept of the store. For example, when you walk into a retail store, you are almost always anonymous. There’s barely any of the personalization you might have if you go to the brand’s website. There are ways that AI can help solve that and make the interaction with the customer a lot more personal. Within the next five to 10 years, almost every company will be using AI in a pretty significant way.
You don’t need to know AI to benefit from it
For an organization to benefit from AI, where should it begin?
I think it should begin with the business problem. It should not start with the technology. It really starts with figuring out what business problem you want to solve and then figuring out how AI can be applied.
If top AI talent is going to companies like Google, how should large organizations meet the AI talent challenge?
Talent is actually the biggest limiting factor for AI right now. There are very few people that can actually build a deep learning model for example. I think it is a two-sided challenge. On one hand, how do you train more people to be able to do these things? On the other hand, how do we make AI easy and useful for organizations? This is where we are focusing. How do we make it so that the existing teams in organizations can leverage AI? Existing developers in traditional organizations shouldn’t have to immerse themselves into the deep learning model in order to benefit from it.
AI today is really where the web and the Internet were in 1994. Everybody sees a lot of promise, but it is still very hard to build upon. Back then, building a web page required a lot of know-how in terms of knowing HTML, for example. From there, we very quickly got to the point where any average business owner could put up a website and enable e-commerce. And that’s where we need to get to with AI. The average business owner should have building blocks that they can put together in an easy way to be able to accomplish their needs without having to build models themselves. That is what we are focusing on – how do we bring AI to many more people, how do we democratize it?
AI still has its issues
What is your view on the black box nature of AI?
In many cases it doesn’t necessarily matter how AI came to a conclusion. The bigger question is about whether the conclusion is useful and ends up creating benefits? Take, for example, product recommendation. If AI recommends a particular product, the way you really test that is whether people buy the recommended product. Of course, there are many other areas where explainability is critical. For example, with medicine or fraud detection it is critical to be able to explain AI’s decisions. That is something where the science needs to evolve a little bit more because these algorithms are fairly generic. Figuring out explainability is a challenge that the field needs to work out.
In your view, what are some of the issues with potential biases resulting from the data that is used to train AI?
That is a big issue. AI is fundamentally not biased, but if the data itself is biased, then it is going to produce biased outputs. That is an area where we as a field need to improve and figure out how we get to the right kinds of best practices. One thing that we observe is that the vast majority of work in building a great model is not actually in the model itself, it is more in the data. Figuring out what data to use, getting the data to the right place, cleaning the data, doing data engineering and feature engineering, and then figuring out what outputs you are looking for and how to use those outputs. That is the bigger challenge in almost any of these problems.
AI @ Google
Google has a variety of AI initiatives. How does the company view AI?
At Google, we emphasize that AI is not something that should be done by only one group. AI is something that should be pervasive. Over the last two years, we’ve gone from around a couple hundred projects at Google using AI to almost 7,500 projects. Sundar [Sundar Pichai, Google CEO] has already said we are an AI-first company. We’ve gone through a rapid transition to becoming an AI company where almost every product utilizes AI as a tool to provide a better user experience. The current level of interest in AI is like nothing I’ve seen before in my career.
Which Google product do you believe AI will impact the most?
AI can have an impact on every Google product. I see AI as more of a paradigm of engineering as opposed to anything else. One of the best examples of AI at Google is Google Photos. All of us have tens of thousands of photos and categorizing them is a painful exercise. Google Photos really took that to the next level by bringing Google-quality search to photos. Translation is another big area. We just publicly shared a demo with the new Pixel 2 smartphone where, if you wear our ear buds, we can translate another speaker using a different language and give you the translation in your ear simultaneously. It is really magical because we can start to bring the world together. Another surprising application of AI for us has been the Gmail smart reply. With email on mobile, we suggest a reply at the bottom of each message. On mobile, that is extremely valuable, because typing something on a mobile phone is not an easy thing to do. We are now finding that over 10% of replies via mobile for Gmail use this smart request feature. Things like that are really showing that AI can not only be accurate but incredibly useful too.
How does AI help you differentiate Google Cloud?
AI is going to be a big differentiator for Google cloud in a number of ways. First is the infrastructure and making Google cloud the fastest place to run AI. Second is the building blocks – the prebuilt models that users can take and customize. We have a wealth of models and knowledge that are there for users to take and use. Third is partners: having partners build on top of this and offer services to customers is really how this can scale. We are putting a lot of emphasis on working closely with partners to make them successful.
What do you say to people that are worried about AI taking over the world?
AI is still a very nascent technology. We are still only a few years beyond being able to take an image and tell a dog from a cat. We still need to figure out how AI can best augment what we do. The big issue for me is one we spoke of earlier – bias. I think that is a more pressing issue than worries about taking over the world. If we are producing algorithms that contain biases, then that becomes a much bigger problem.
We are really at the beginning of a journey for something that is going to have a positive impact for people everywhere over the course of the next 10 to 20 years. It is up to us to figure out how best it will have an impact. I’m really excited about having this opportunity to really shape the next generation of technology.
Babak Hodjat is a Co-founder and the Chief Executive Officer of Sentient. A serial entrepreneur, Babak has started numerous Silicon Valley companies as main inventor and technologist.
Prior to co-founding Sentient, he was senior director of engineering at Sybase iAnywhere, where he led mobile solutions engineering. Babak was also a co-founder, CTO and board member of Dejima Inc., which was acquired by Sybase in April 2004. He is the primary inventor of Dejima’s patented, agent-oriented technology applied to intelligent interfaces for mobile and enterprise computing – the technology behind Apple’s Siri.
He has 21 granted or pending patents to his name and holds a PhD in Machine Intelligence from Kyushu University, in Fukuoka, Japan. Capgemini’s Digital Transformation Institute spoke to Babak to understand specific applications of AI across industries.
Building disruptive AI that delivers today
Can you tell us a bit about Sentient and your background?
I’m the co-founder and CEO of Sentient Technologies, where we scale AI to disrupt industries. We build products that disrupt industries and we’re organized into three business units. One is in asset management, investment management or trading. The second is ‘intelligent commerce’ – the AI-enablement of the shopper journey. The third is digital media optimization – effectively building websites that adapt themselves to different users.
I have a PhD in Machine Intelligence from Kyushu University in Japan. In a past life, I set up a startup and I was the main inventor behind the National Language Conversational Technology, which ended up being the underlying technology behind Siri. A big chunk of my team went on to work at Siri, then became part of Apple, and now we have started this.
Are you already seeing some early benefits where you have deployed Sentient AI for clients?
Clients don’t put money into this because it’s some sort of ‘magical AI’. For example, when we sell AI enablement for retailers, we are measured by indicators such as the increase in conversion rates or the average order value. Our business model is based on that approach and we are incentivized by the performance of our AI-enabled products. The same thing goes for digital media optimization. The user defines the measure – such as conversion rates online – and that’s what we’re optimizing, and we show where we actually improved on that. AI is doing that, but how we get paid is a function of that improvement.
Can you tell us more about how AI delivers benefit in commerce?
E-commerce has really not changed significantly in the past decade or more. If you want to buy something online, your experience is more or less the same. You go online you see pages upon pages of seemingly randomly ordered images of what you’re looking for. You then have to scroll through it until you find what you want, or you have to put in lots of filters and so forth. It’s cumbersome, difficult and it’s been a failure. The reality is that online conversion rates are at around 3%, which translates to a 97% failure rate.
So how do you change that? An AI-enabled shopping assistant interacts with you and through that interaction, you can quickly find exactly what you’re looking for. We took AI and we had it learn what there is for sale so that it has a full understanding of everything that this e-commerce site offers. The AI then sits in-between that understanding it has developed and the user. Within a few clicks, users get to exactly what they’re looking for.
We have deployments for example with Sunglass Hut or with Skechers, and in these cases, we have seen 30 or 40% improvements in conversion rates. Not just that, we’ve also been able to increase the average order value which is an indication that the users are actually finding what they’re looking for and therefore, the order value is going up. There’s this anecdote that users at any given point in time, are only seeing about 10% of the inventory. Post our AI deployment, 75% to 80% of the inventory is actually seen by the users because the users are directed and are helped by the AI to navigate the sites.
Can you explain how AI-based trading works?
The AI makes all the decisions from what instruments, how to trade it, whether to go long or short, how long to hold, how much to buy and when to cover or sell. All of these decisions are made by the AI. The AI-based desks have certain risk and return profiles. That depends on the appetite of the investor in the fund.
The AI can very much optimize for different outcomes. These are problems that we call multi-objective problems, which means that we’re actually solving for more than one objective rather than only solving for making money. The AI is solving for making money while reducing risk and, by doing so, there are multiple solutions to a multi-objective problem. You could have a high-risk but high-return solution that is acceptable for a certain class of your customers, and you could have another solution that is much more conservative with, of course, lower returns. So, when you give the AI a multi-objective problem, rather than coming back with a single solution it gives you a spectrum of possible, acceptable solutions. Based on that, you can actually market different offerings depending on the appetite of your customers.
Getting AI off the ground
How do you measure the ROI from an AI deployment?
In the case of digital media or intelligent commerce, it’s pretty straightforward, because organizations already have the cost of acquisition for a user. So, when a user hits a website in e-commerce, every click is a cost and bringing the user to the page is a cost. So, the larger the percentage of people that convert, the less the cost of conversion. And, of course, the increase in conversion – or an increase in Average Order Value – mean more revenue for the organization. We measure that improvement as reflected in dollars. In the case of digital media optimization, there’s also the traffic. Here, our business model is currently called a traffic-based business model. Depending on the traffic on the site, you have a different value to that site. So, if we can improve the value of the conversion relative to the traffic, then we can improve on that particular KPI.
In trading, it is very straightforward and just like any other hedge fund. We’re measuring consistency of returns through time and it depends on the appetite for risk with the different desks. Our model is very similar to a hedge fund model, where there is a commission and percentage of the returns. It’s commonly known as the 2-20, but it could be variations of that, so 2% management fee, 20% of the returns.
How do you train the AI and how much time does it take?
Building an AI product is not trivial. It takes a lot of effort and work and we’re getting better and better at it. Today, it’s a matter of weeks to get from signing an agreement to actual deployment. The actual custom work is maybe 5% or lower, because for most of these categories that are being sold online, we already have models. The whole thing is very much industrialized so that we can manage multiple customers at the same time as on-boarding them and taking new inventory and changes on a daily basis.
How should a large organization go about deploying AI?
I don’t think of AI as a one-size-fits-all kind of a solution. I don’t think a big organization would ever say, “We’re going to standardize and AI-enable everything”. This is because the AI-enablement for different aspects of the organization is going to manifest itself very differently. So, you’re going to have multiple AI-enabled products, each one applied to a different aspect of an organization. There are companies that are starting to think strategically about AI. The usual way is to identify the two or three different areas where it’s fastest to deploy an AI solution and then move to a wider-scale deployment.
A bright future for AI
Which sectors do you think are conducive to AI?
We believe that AI has very broad applicability. For example, there’s a lot of AI being applied in healthcare these days. but not too many people looking at agriculture. But we have been working with OpenAg at MIT Media Lab and undertaking some projects where the AI would discover recipes for how to grow plants and environmentally control situations. For example, we see that the AI is discovering how to grow plants – deciding which spectrum of light to shine, the humidity, water levels, and minerals to be added. You can actually optimize for taste, believe it or not. I wouldn’t say there’s an industry in which AI is not right for disruption.
How do you see the future of AI?
I think AI is going to become more and more ubiquitous in products across many different industries. We will see a lot of AI though in many instances we will not recognize it as AI. It’s just going to be natural – you will interface with it without knowing it is AI. I think AI has a bright future in things like the IoT, in agriculture, healthcare, and across many different enterprises such as banking and insurance. AI will also manifest itself in what we more commonly know as interfaces for robots. Robotics is an industry that’s very much hardware-originated and the interface, therefore, has always been kind of neglected. I think we will come to a point where the best interface to robots are going to be conversational and augmented so that the robot can actually perceive its surroundings. There is a very bright future for AI.
Rob High is an IBM Fellow, Vice President and Chief Technology Officer, IBM Watson. He has overall responsibility to drive Watson technical strategy and thought leadership. As a key member of the Watson Leadership team, Rob works collaboratively with the Watson engineering, research, and development teams across IBM. Prior to joining the Watson team, Rob was Chief Architect for the SOA Foundation and member of the IBM Academy of Technology.
Capgemini’s Digital Transformation Institute spoke to Rob to understand how AI is permeating across organizations of all sizes and how it is being used.
What will be one of the biggest changes that AI will bring to everyday experiences?
This goes to some basic fundamentals of human–machine interfaces. If you look at the history of computing, for the last 70 years of modern computing humans have essentially had to adapt to the constraints of the computer. The human–machine interface has always been defined by constraints in how the computer is capable of expressing itself, from punch cards in the early days to sliders today. They are still largely defined by forms of interactions that are not natural to human beings. So, one of the benefits that AI can bring to the table is a much more natural form of interaction. If the AI is already able to understand our language, what we hear, and what we see, then the next logical step is for us to begin to improve the way that it interacts with us. This is an area that will be most interesting to watch when we think about the application of AI in solving real-world problems.
Helping people be better at what they do
How can large organizations benefit from AI?
The primary benefit comes from helping people who do things today, do those things better. One practical way is where AI helps you consider more sources of information in a world where the amount of information that we are exposed to is growing exponentially, and far exceeds the capacity for us to possibly consume and assimilate. If you are an organization that has a very large call center, for example, where you are having to deal with very large numbers of customers who come to you for support or advice, then having a cognitive system can help significantly. For instance, you can augment call center staff by offloading some of the more mundane tasks. Call center advisors can do a better job in answering clients’ questions and concerns by finding information that would otherwise be difficult to access. Cognitive systems can also help companies where there is significant diversity of expertise across the organization. For instance, doctors who have been in practice for 20 or 30 years will have built up a lot of expertise from seeing many more cases and variations on those cases in different patients. This expertise can be captured and democratized through AI.
Which industries are best-placed to benefit from AI?
Currently, the healthcare industry recognizes the need to get this kind of assistance for their cognitive processing. Doctors know that it is very difficult for them to keep up with the very latest advances in their domain. They have therefore reached out very aggressively to adopt systems that make better treatment decisions or improve their ability to diagnose. Among other industries, financial services definitely has an interest. For example, financial advisors spend a lot of their time researching different investment choices for their clients, and that is a process that requires a lot of reading. It requires going through a ton of information. The more that they can get assistance from the cognitive system, the more information they can assimilate to make better decisions for their clients. Retail is another industry where I believe AI can significantly enhance customer interaction.
Get the use case right
Where can AI provide most benefit? Are there specific categories of data where AI can have greater impact?
One of the unique characteristics that AI brings to the table – which traditional computing doesn’t – is its ability to look at, and understand, human expression. By this, I mean things like text or literature and audio and video. To date, this data that has largely been the unique domain of human expression, and this is where AI has a tremendous opportunity to make a difference. In general, any place that has qualitative information that, up until now, has largely been ignored because of the limits of computational systems to understand it.
How should organizations go about implementing AI?
I believe it is always important to start with the use case, because the use case is going to then tell you how much work you have to do to get ready and also something about the value that you will get from having invested your time and energy. When you have a use case that you are ready to invest in – and it is well defined, scoped, and targeted – then you begin to look at the most effective way of evolving through that use case. You start to think of the minimum viable product that you need to create in order to get started and yield some of the benefits. You start there and grow. But if you don’t have the use case, and you don’t know the value of it or the market conditions associated with it, that makes it hard to answer any other questions.
Most popular use case: the contact center
What do you believe are the top use cases for AI?
The most popular use case right now is in the contact center. There is just so much room for improvement in the customer experience, the abilities of the agents involved, and the relationship between the institution and its customers. I see that as being an area of massive growth in the world of AI. It dramatically transforms the contact center industry, but more importantly the relationship between institutions and their clients.
Bot first, or human first?
If it comes to deciding whether to deploy a bot, or a human, how should an organization go about it?
There are a couple of different schools of thought. One is that you always send your clients to the bot to begin with, but build into the bot the ability to detect when the user wants to opt out and go directly to human. This approach is similar to the evolution I think we have seen in the IVR space. Many institutions are now forcing you to their IVR, and then in the IVR there is an option to talk to a human. The other approach is to make that choice available as different entry points into your organization. So, for example, you could have a web page where you are selling a product, and you have a catalogue of all the products in your store, and you deploy the ability to both chat with a bot or chat with human being there and let the customer make their choice.
Some of that has to do with the nature of the customer base you are dealing with. For example, there are a whole category or people in certain parts of the world that prefer to deal with a chatbot. To them, it seems anonymous and also gives them freedom to ask what they believe are silly questions and not be judged by a human being on the other end. Older people under certain circumstances may prefer to talk to a human being. So, by giving them that choice, people can go for the option that is most useful and comfortable for them. The third case is when you always talk to human beings. In this case, you really need to give AI to the agent and help them do a better job serving their clients. The idea is to make their client service consistent, more efficient, and with better quality of information.
You don’t need to be an expert in AI to benefit from it
If finding AI talent is a significant challenge for many organizations, how should companies develop AI capabilities?
I don’t think you need to be an expert in the technology of AI to benefit from it. You don’t need to be skilled in AI, you need to be skilled in your use case. Now, that is not to say there aren’t going to be cases where you need to build your own algorithms. But I think there are a lot of things that people can do with AI services that enable them to support the use case without having to know all the details of how that AI works.
How do you see AI impacting on jobs in the future?
I think that there will be some displacement of jobs, but I think for every job lost, there will be many more jobs that are gained. The role of AI is not to replace humans, it is to augment humans. It is about helping us be better at what we do; it is not about doing what we do. It means that in every job that is out there, where there is room for improvement, there is going to be a room for AI to improve the job, and make people better at what they do. So, it will be incumbent upon people to determine for themselves whether they want to take advantage of the benefits of AI. If they refuse, they are going to be at some risk. But for everybody else that want to improve themselves, they are going to find that AI actually makes their job better and easier to perform at a higher level. In the end, it is really about how we augment human intelligence, not about how we replace human intelligence.
Lili Cheng is a Distinguished Engineer and Corporate Vice President of the AI and Research division at Microsoft. Lili is responsible for Microsoft’s AI Developer services such as Bot Framework and Cognitive Services. Lili founded the Social Computing Group in Microsoft Research & Future Social Experiences (FUSE) Labs, was the Director of User Experience for Microsoft Windows, and innovated on technical infrastructure in the areas of real time data and search.
Prior to joining Microsoft, Cheng worked in Apple Computer’s Advanced Technology Group on the User Interface research team, where she focused on QuickTime Conferencing and QuickTime VR. She has also taught design at NYU and Harvard University.
Capgemini’s Digital Transformation Institute spoke to Lili to understand her views on AI and what it means for large organizations.
Experiment, experiment, and experiment
How can AI help large organizations?
We believe there are two broad areas. One is about B2C and the other is about how people get work done inside a company.
Let’s start with the first area. For example, if I want to interact with my insurance company today, it is still pretty hard to get basic help for some of the tasks that I want to do. One area we have been focusing on therefore is customer support: helping companies answer both simple and more complex questions and working with them to improve the way they reach their customers.
The other area is how people get work done inside their companies. We are focusing on how people work on projects, manage email, schedule their meetings, communicate via messaging, and so on. We believe AI can be of significant help here. Take this very discussion for instance. We should be able to have a more dynamic way of understanding the context quickly. I need to go into my mail, find the right email, and quickly look at a couple of other things before I come into the discussion. Then, we have to figure how to record this interaction and take notes and follow up. People spend so much time in meetings and it it’s an area of dissatisfaction for many because they often feel it’s not the best use of their time. We are thinking about how to improve the intelligence when people come together in groups.
How should large organizations approach AI?
The rate at which companies and researchers are sharing and open sourcing their technology is really unprecedented. It is great because it means you will have more people who actually have domain expertise and can author and create intelligence. We don’t want people to have a PhD in AI to benefit from it. The tools need to be usable by people with different skill levels. The person actually using the tool is probably one of the best people to design it.
What are some of the areas where AI can have the most impact?
The areas where we think companies will want more intelligence is in managing the financial data flow – who your customers are, what they are doing, who has paid their bills and so on. Customer support is obviously a big area and another really interesting area is time management. We recently acquired a start-up called VoloMetrix. They look at workplace satisfaction: how people spend their time; how satisfied they are with their job; how much time do they spend on meetings and email and what is the best way to help people be more proactive and spend their time more meaningfully? In a certain sense, computers have taught people to be very distracted. One of the things that could therefore be really interesting is how AI can help people manage their time more effectively.
What should organizations bear in mind when they deploy AI solutions?
The one thing you need is simply better data. Having the data so that you know where you are and the top three scenarios for where you want to be.
Being able to experiment quickly with some customers before building the entire system is also really critical. This is because if you want the perfect AI solution it can take a while to get all your data, integrate it and roll it out. Also, there’ll be times when the things that your customers want are going to change. A year ago, for example, people were a lot less interested in speech. Today, I think people feel the need to not just have text-based solutions, but also speech-based solutions. So, I think being able to experiment quickly, and have an experimental group that can try things out ahead of the entire organization will be crucial.
How should organizations decide what should be driven by a bot versus what should be driven by a human?
When you first roll out your system it is not going to be very smart because it doesn’t have that much data. In the beginning it’s therefore particularly important that you have people to override, train and manage the system. But over time, as the system learns, then we can project with high probability that this task can be automated. At that point you can automate it and your people can focus on complex tasks, new areas, or things that are more ambiguous.
We are better prepared for the onset of AI than any other technology in the past
AI will likely impact on jobs. How do you believe organizations should approach it?
I think that organizations should know that their people are probably their most valuable asset. Without your employees and your customers, you don’t have a business. So, making sure people’s jobs are interesting is really critical. I also think it’s really important to be aware of why you are automating and which new opportunities you need to go after. We have been focusing more on areas that are still fairly hard to automate and, so I don’t think that we are seeing job displacements. But I do think that it’s something that companies and governments need to work on together to ensure we all understand the impact.
What do you see as the future for AI?
It’s important to remember that every technology has changed the workforce. If you think of the Internet, the ability to access the world’s information instantly has changed a lot of jobs. With AI, compared to other technologies, we are probably more mindful of the impact on how work may change. I think it is our responsibility to make sure that it is augmenting how people work. We need to ensure that people can spend their time more effectively and we are really committed to that. We are all thinking a lot more about this than we thought about, for example, the Internet. We didn’t have these conversations when the worldwide web was emerging, because I don’t think people really had the foresight to think about how it was going to transform the way everybody works and lives. This time around with AI we are a little bit smarter to be asking these questions earlier.
Amr Awadallah is the Chief Technology Officer of Cloudera. Before co-founding Cloudera in 2008, Amr (@awadallah) was an Entrepreneur-in-Residence at Accel Partners. Prior to joining Accel he served as Vice President of Product Intelligence Engineering at Yahoo!, and ran one of the very first organizations to use Hadoop for data analysis and business intelligence.
Amr joined Yahoo after they acquired his first startup, VivaSmart, in July of 2000.
Amr holds a Bachelor’s and Master’s degrees in Electrical Engineering from Cairo University, Egypt, and a Doctorate in Electrical Engineering from Stanford University. Capgemini’s Digital Transformation Institute spoke to Amr to understand how organizations can benefit in their decision-making using AI.
How would you define AI and what level of take-up are you seeing?
I prefer to see AI as the automation of decision-making as opposed to ‘intelligence’. It’s about learning about multiple areas and then making decisions like a human does, but in a much quicker, faster, more reliable and more accurate manner. So, it is really about the automation of decisions.
We have more than 1,000 customers worldwide. But I would say only a handful are truly undertaking compelling AI implementations and automating decisions in a unique way. Most companies are still just figuring out how to collect, sort and catalogue their data.
Make sure there’s a problem that AI can solve
Do you think organizations understand AI’s potential?
I don’t. Many organizations are very excited by the buzz but don’t really understand what it means. It is important to understand the limitations of what can and cannot be done with AI. A significant reason for the confusion is the name ‘intelligence’, because it is not really about intelligence. It is actually about process and decisions. It is about learning how certain processes and decisions work within the organization and then automating them.
What does it take for organizations to deliver benefits from AI and make a success of their programs?
The first imperative is making sure you have the proper systems and the right skills. The second is ensuring that you identify the initial use cases that can benefit from AI, as opposed to just adopting AI for the fun of adopting AI. You need to have a very well-defined use case or a very well-defined problem that AI can help solve.
The benefits are there, but realizing them requires hard work
What differentiates the companies that benefit from AI?
What differentiates them is the first-mover advantage. Those organizations that are reaping the benefits today started thinking about AI up to eight years ago. They began building the proper foundations from a data platform perspective. They started training their people in the skills required. Now, eight years later, they are at the point where they are truly able to leverage AI.
There is a kind of basic hierarchy of needs or processes that organizations need to go through. For example, they need to be able to collect data and different types of data across their organization, including real-world data. So, having more sensors and technologies that bring in that data from the real world is the top priority. Number two is processing that data. Being able to manage, process and clean it up at scale is still something that most organizations struggle to do. Then you start with basic analytics, gradually moving to advanced analytics and then AI. The last step is the most advanced, which is to automate. Depending on the organization, this stage can take anywhere from two years if they are very fast or as much as eight years.
What is holding organizations back from embracing AI?
It is a combination of two things – a proper data infrastructure and skills. The automation of decisions is dependent on having massive amounts of data that you use to train the algorithms and undertake the automation. But most organizations don’t have very good hygiene levels when it comes to data management. In terms of skills, the skills that most organizations have are very focused on business intelligence, but it’s a discipline that’s very reactive in nature. Here, you are automating reports that a human needs to read to make a decision. That is the very different skill set of automating decisions, which requires training to bring people up to speed.
What should large organizations do about skills?
One thing is clear – large organizations are not going to be able to just hire people who are good at AI, because they are very hard to find. The best thing to do is to train your existing people. Another way of working around the skills challenge is to develop an AI ecosystem, where a wider spectrum of companies can tap into the power of AI.
Every decision is ripe for AI
Are there specific sectors where you believe AI will see greater adoption?
We are seeing across-the-board adoption. Typically, you have more agile, innovation-ready like finance and telecoms, but we are seeing appetite in many sectors. We are seeing it in power, smart grids and utilities. We are seeing it in manufacturing. We are seeing it in shipping, for companies that manage big fleets of ships across the world. Multiple sectors are very receptive and understand the implications of AI.
Are there particular business functions that are more suited to AI?
Any function where humans make repeatable decisions can be impacted by AI. It is applicable to a lawyer reviewing a contract to a doctor making a diagnosis. Another area is prediction. For example, predicting ahead of time the failure of a piece of equipment, such as an elevator or power generator, by analyzing the patterns of data emerging. Predictive analytics and maintenance is a very common underlying theme and gets classified under the AI umbrella. Another function is anomaly detection – trying to detect weird behavior or patterns within an organization. For example, a bank trying to detect patterns to counter money laundering.
The future for AI
How long do you think it will take for AI to become mainstream?
In the enterprise software arena, as opposed to consumer electronics, everything takes time to mature. It is not going to be like smartphone adoption. Based on historical experience, I would say five more years.
What is your view on the impact of AI on jobs?
In my view, AI is not replacing jobs. It is replacing decisions that certain jobs make. Take the example of lawyers I provided earlier. You will require less lawyers, but you will still require lawyers. We will need lawyers to review the exceptions – the challenging cases that the AI system cannot automate. Any new technology introduces new levels of efficiency and we as humans have to adapt. This was true with the industrial revolution. When we created the steam engine and the electric engine, a lot of people had to re-train. There will be a lot of efficiencies that will have a significant impact on many jobs, but jobs are not being completely eliminated.
What are some of the skills that people will need to survive in the AI age?
Firstly, skills around maintaining, operating, training and overseeing AI systems to ensure they are working correctly. The other important skill that will be required across all sorts of jobs would be the willingness to learn new skills. That is because expectations will change significantly as machines start to take on more and more decisions from humans.