A different story of AI project structuring
In the news we read that AI is something for the Broad IT-developer market. Why would that be?
We believe that AI and ML are essentially designed as statistical tasks with special optimization potentials and a special kind of process engineering as an influencing guidance task in the AI-project.
How long an AI project lasts does not depend on the correct choice of the AI-Method, but on the usability and relevance potential of the basic data for the target question.
Therefore, it is important that you do not ask at the beginning of the trip in an AI / ML project what the AI / ML can do, but I know how the needle should look like in a haystack? The IT integration is completely subordinate here. Learn the Basics –> Statistics and Optimization in Statistics.
But what about a project in AI / ML? How do you structure the problem?
The classic IT-PM approach set in the agile world:
Approach to AI / ML procedure:
- Identify the need and motivate the customer
- Form a hypothesis based on the basic data with the aim of the customer’s request.
- Show with the proof of the hypothesis that it is possible to use AI and ML.
- Create a pilot / PoC Situation
- Realize the project to optimize the pilot and make it an executable product (model with ongoing optimization).
About Design Thinking of the idea for Lean / Agile of the project until the ongoing optimization of the product. The lifecycle contains everything you have seen in the past 30 years of IT, the horizon of development.