AI projects are not IT projects
You might be relying heavily on code, developers, cloud services and data stored in systems managed by IT but an AI project is a business project, the focus should be about solving a problem, not investigating the technology.
Momentum is important
Momentum makes these projects successful. You’ll need to be bold and disruptive to make progress. The best leaders give people the freedom to be creative, find one of these to be your sponsor.
Building the team is vital
Good AI projects enable a team not replace a team. For success you should use your existing experience to find new results across larger data sets, opening up scale that never existed before.
I use this equation to build out the team
Success = Platform + Expertise + Data + Domain Knowledge
Platform = Extensible, secure and open public cloud
Expertise = Developers, data science, cloud architect
Data = Organisation or 3rd party source. Note you don’t need it all to get started
Domain Knowledge = Field expert
Success should come quickly
If you have put the right team together you should be getting results or proving the validity of the project in just a few weeks. In IT we call these ‘sprints’. One or two ‘sprints’ and you should be making progress. I’ve seen results much faster than this too.
Investment should follow success
Finally, major investment should not be provided until success has been shown, but it should then come quickly. Losing momentum dampens enthusiasm, causes rot to set in and then projects fail. Failure is not a bad thing, it helps us learn but failing to move a successful project forward because it has nowhere to go means we’ve got the wrong sponsor or are not solving the right problem. Reset and go again!