Failing on delivering on a project while spending all the budget with mediocre results to show is one of the nightmares of your career path.
Lots of painful “lessons learned” can be avoided if AI projects are setup the right way from the beginning
We would never wish that scenario to anyone. Nevertheless, it happens often during high-risk artificial intelligence projects.
The most common scenario is the following: At the beginning, people are super excited and hyped by all the news around AI and the potential for lots of PR and success.
First few months flow OK and the project seems to be going well, although on the AI side the results are not there early but this is considered normal as AI is R&D.
Few months down the road it goes like that:
1) The AI is near complete and first attempts for deployment are being made
2) AI behaves erratically and there is lots of disappointment that initially is managed up and down
3) AI team cannot really deliver and then project starts prolonging. More budget is required
4) Trust is lost in the AI team. Extra people are being brought in. Some of the previous AI people quit
5) Lots of churn and burn and the project is at high risk
6) Project is considered a failure and there is lots of finger pointing and responsibility cascades
7) Managers and organization become risk averse and avoid AI projects.
There is no easy way out of this. Only experienced teams in the development and deployment of commercial AI projects will survive this early era of high expectations and hype and be able to deliver projects that create value.
We encapsulated our experience in a questionnaire that we always share with our clients before we engage in a project. It helps both parties understand the level of effort, risk and potential for value out of the project. It aligns the teams for success and helps manage risk early while developing the AI system.