Skip to Content

Decoding Artificial Intelligence

October 18, 2020

AI is like a magic trick. While the word “artificial intelligence” could make us think that robots have their own mind, its technology and application can tell a different story.

Sure, AI development involves a lot of computer programming, creating and iterating algorithms, and database building. But up until 2020 at least, AI applications have mostly been designed to complete unwanted tasks rather than replace people’s jobs (or humanity). In fact, we might just need to learn it right to leverage the most out of AI.

Application of AI within business

Using one of the real examples applied to Peloton Innovations, an AI Canvas (Figure 1) is developed to improve home security using AI. By combining historical and real-time data, the model is designed to predict, judge, act, and evaluate tasks based on the client’s needs.

Ideally, clarifying these critical factors within the process of implementing AI into home security would either reduce the business’s costs or enhance performance. With its increasingly accurate and high-fidelity predictions, it will be able to predict home intruders before they even enter.

However, the greater potential may also mean greater risks. As Gartner predicted, 80% of AI projects through 2020 will remain alchemy, and frontline workers will find it difficult and ineffective to implement the “upgrade” as they used to. This is the challenging part for business decision-makers, as they will need to have clarity on what AI will contribute, its interactions with human workers, and how it will be used to influence decisions and measure success. They will also have to decide on the types of data input to train, operate, and improve the AI model. It is advised that “to get started with AI, your challenge is to identify the key decisions in your organization where the outcome hinges on uncertainty.”

AI talent

AI projects generally involve difficult and complex decisions, with high involvement of data input and navigation. Fundamentally, this requires a lot of technical experts to maintain and sustain the background running of the technology, but there might be more to it than that.

Senior technology leaders have suggested that while computers are outpacing human counterparts in performing repetitive tasks, there are also new portfolios of jobs that are very dependent on humans: junior threat hunter, analytics translator, and conversation designer. Junior threat hunters need clear and concise communication skills to investigate and evaluate unusual activities and threats on networks. Analytics translators need to synthesize and put data into a particular business context, acting as a bridge between the technical knowledge of data scientists and the operational expertise of managers. Conversation designers, for example, for a chatbot, require experts in conversation and personality creation to write scripts that map out user experiences.

As AI-related technology continues to develop, it sure will need a lot more people to work on not only the cognitive aspects but also the talent to work on its humanity-related aspects and enhance its user experience, especially when a lack of diversity in AI is the main issue to be tackled. A 2020 AI Talent Report points out that women constitute a mere 15% of Facebook’s AI research staff and 10% of Google’s, and that only 4% and 2.5% respectively of them are African Americans. Negative effects such as biases and discrepancies towards minorities within programs and systems designed are then likely to be presented and applied.

AI in the Future

That said we are still at an early stage of AI development. A report from the Capgemini Research Institute shows that only 13% of businesses have successfully deployed use cases of AI in production and continue to scale more throughout multiple business teams – of which retail and life science are leading the scaling race across sectors. The recent economic shutdown due to COVID-19 is expected to affect business performance and suspend investments in AI initiatives.

But with risks, there also come opportunities. The lockdown has made working from home (WFH) the new norm and increased people’s dependence on shopping online. It will be interesting to see how businesses in the retail sector will continue to leverage AI in capturing gaps in markets and optimize their presence online.


Emily Suet-Ching Lee

Senior Analyst – Enterprise Apps

Capgemini Hong Kong