Most organisations that start with AI do not have too few ideas. They have too many, and no fair way to compare them. The result is that the idea with the most enthusiastic sponsor gets built first, whether or not it is the right one.
Five questions that sort ideas quickly
- Business value: if this works, what does it change in revenue, cost or efficiency?
- Data readiness: does the data already exist, and can we trust it?
- Feasibility: is the technique proven, and can it connect to our systems?
- Risk and compliance: what happens when it is wrong, and can a human stay in the loop?
- Adoption ease: will the people who must use it want to?
Agree the weights first
Before scoring any idea, stakeholders agree how much each question matters. Doing this first stops the scoring from being steered toward a favourite.
What the score is for
The score is not an answer. It is a structured way to start a better conversation, and it makes the trade-offs visible: a high-value idea with poor data readiness may belong in the “later” column until the data is fixed.
[Expand this post with your own examples from real projects.]