If you don't have the data available, implementing an AI use case becomes a data

If you don’t have the data available, implementing an AI use case becomes a data gathering death march, often crossing organizational boundaries. Instead of spending 80% of the project time on building the use case, this time is spent on building the prerequisites. Overall project times increase and projects might remain unfinished.

If you don’t have the data available, imagination of what can be done can fall short. Good use cases might remain undiscovered. Instead of people experimenting with data to find possible use cases, this becomes a theoretical exercise.

If your data doesn’t have the right quality, your AI implementation will fail. Ex falso quodlibet. Garbage in, garbage out. Another sunk project.

With the rise of GenAI, AI has been democratized. More and more roles in your team can build AI use cases. Not having the right data basis is therefore increasing opportunity cost. Instead of “just” a small team of AI experts failing their projects because of missing/bad quality data, many teams in the company are failing to deliver their GenAI projects.

To be clear: I’m not saying to hold the entire company back to build a full-fledged data platform if you haven’t done so yet. No. But you definitely need to get the vision right and prioritize building the data foundation. From my experience though, building a data platform can easily derail into a huge program failing after a long project time. The key here is to decide the architecture vision and building blocks early and then build in iterations, making sure that each iteration is delivering business value.

Happy to help!

What are your thoughts on this? Join the discussion in Itumeleng’s initial post - I wasn’t able to fit my view in the character limit of a comment, hence the repost. Maybe we also get more people joining the discussion.

---