From pilot to production: why most AI projects stall

Many organisations have now run an AI pilot. The model works, the demo gets applause, and then nothing else happens. In our experience, most projects stall not because of the technology, but because of everything around it.

Three things need to be in place

To put AI into production, you need three things in place: data that comes in reliably, a team that can maintain the model, and a process in which the output is actually used. If one is missing, the project remains a demo.

• Data that comes in reliably

• A team that can maintain the model

• A process in which the output is actually used

In this article, we look at each of these, with examples from projects for local authorities and in healthcare.