Deploy successful AI strategies // Part 2: Business Case

Usually, existing enterprise data offers a veritable Eldorado of opportunities for a wide variety of use cases for the application of machine learning or artificial intelligence.

So where to start?

The worst possible approach is to first set up an AI model and then look for a suitable use case. It’s better to know about your customers’ problems and needs first. Learn as much as you can about the difficulties, challenges and problems of your target audience. The bigger the identified problem is for your customer, the easier you will be able to convince them of your solution. However, outside the ivory tower, professional life is often much more complex and insights into your customer’s needs are only partially available (or worse: not possible). In these cases and from a business perspective, what you need is a Plan B so you can start to identify business cases: Increase revenue or reduce costs (or both). Every economically active organization has to deal with these two metrics on a daily basis and therefore they are definitely relevant for them. From this starting point, you can derive three business case areas for your endeavour:

1. Development of new business models or products and services, based on existing data sources (revenue increase)
2. Extension and optimization of an existing product and service catalog with AI functionalities (revenue increase)
3. Optimization of business processes (cost reduction)

Once you have identified a suitable business case, you can consider to what extent AI models represent a solution. In addition, check whether the business case actually requires AI support, or whether this is cracking a nut with a sledgehammer. AI projects and resources are expensive and the use of AI is not always the most target-oriented. Depending on the industry or company size, the identification, evaluation and classification of use cases is a collaborative effort between business departments and IT or data scientists and data engineers. Only very few companies will be in the fortunate position of having both the technical knowledge and the data expertise in one and the same person or organizational unit. It is therefore important that all areas achieve a good result here on an equal footing. The functional areas must be able to demonstrate the economic added value, whereas the data jugglers should assess the feasibility and any implementation risks.

 Key take aways

• What added value does a use case deliver for my customers or the company?
• Is an AI solution suitable for the use case?
• If applicable, which projects are to be implemented already in your company and what's their use case?
• What criteria are used to identify or evaluate new use cases?

Return to overview/introduction