Skip to Content

Maximizing ROI across the three components of AI enablement

Capgemini
13 May 2020

Artificial intelligence (AI) has become a must-have for organizations looking to gain or extend competitive advantage in today’s complex markets.

AI solutions can provide immense value to an organization. However, scaled implementation of these solutions has proven very challenging for a variety of reasons. Variables such as data quality, data transformation, use-case prioritization, talent, and governance all play a large part in the AI value equation. Overlooking even one variable in the AI-enablement equation can bring significant hurdles.

IDC has found only 35% of organizations get AI projects into production successfully. When we add this to the unprecedented global economic disruption, AI implementations need to be extremely well planned. Through proper planning, maximum ROI can be realized with payback periods that are in weeks or months, but definitely not years.

So how can organizations formulate or refine a plan of attack in today’s market? In a series of blogs, I will answer these and other questions. In summary, though, all AI projects must include planning and alignment across three fundamental components (e.g. three legs of the stool): business use cases, data, and governance.

Business use cases

Particularly with the current global crisis, it has never been more important to leverage AI to help predict the future and prescribe actions that will mitigate risks. From enhanced sales and revenue forecasting inside of our “new normal” business operations to predicting how best to open up micro-economies after an outbreak, we all need to move beyond life in a rear-view mirror. We must take advantage of AI now and forever.

Across functional areas such as customer experience, employee experience, supply-chain optimization, and Internet of Things, the first task is to cast the net wide to find potential AI use cases. By identifying a larger number of use-case candidates (e.g. 40 to 50 instead of five to 10), we are more likely to find the low-hanging fruit, which I define as having the following characteristics:

(1) High value
(2) Low cost
(3) Data readily available to solve the use case
(4) An identified executive sponsor who is very supportive.

The opportunities for AI to improve business performance are wide-ranging. With the right data, value can be extracted from any business unit. Speaking of data…

Data

At its core, AI is fueled by data. People say data is the new oil and I agree with this to some extent, but the oil is really the AI data applications. The data itself, in most cases, is not independently valuable.

Yet quality data is critical to any AI program. From data acquisition and ingestion to transformation and distribution, organizations must constantly improve their data quality to keep the AI train running. However, mistakes have been made whereby organizations think of data transformation and AI-enablement as relatively independent efforts. Aligning your data-transformation roadmap with your AI-enablement roadmap is critical to maximizing success. In fact, if these two component are aligned properly, self-funding of an entire AI program is very feasible.

Governance

So you’ve identified and prioritized your AI business use cases and have an aligned roadmap for the underlying data transformation. You should be in good shape, right? Wrong. Above all, good data management – and, by extension, good AI strategy – calls for sound governance. Unplanned legal or compliance issues can not only hamstring AI efforts, but cost your business dearly in punitive measures and eroded customer loyalty if not addressed properly.

For example, customer personal information should not be used without permission. Most understand this is data-governance priority number one. However, it’s just one of many roadmap items that must be incorporated into an organization’s strategy. Governance plans must address:

  • What is the role of IT in data, analytics, and AI?
  • What is the role of the business?
  • Who owns the data (e.g. customer, product maker, vendor, etc.)?
  • How should our business glossary change?
  • What should we do to improve our data security and compliance?
  • How should our data, analytics, and AI organization be structured?
  • How should we manage partnerships and alliances during these ever-changing times?

The fundamental objective of any AI journey is to drive maximum value for the business, while improving data quality and minimizing data costs. No matter your industry or geography, a well-defined AI strategy that has proper alignment across use cases, data, and governance should form the backbone of your business. This is particularly important now given the more unpredictable future caused by events of the past few months.