KIST INSIGHT | CARL KIRBY
Why AI Should Start With a Business Problem
The question is not “Where can we put AI?” It is “What are we trying to improve?”
Many businesses know they should be doing something with AI. That pressure creates a predictable mistake: starting with the technology and then looking for somewhere to use it.
Carl Kirby’s approach at KIST Consulting reverses that sequence. Start with the business problem, the process or the capacity constraint. Understand what is actually happening. Only then decide whether AI belongs in the answer.
A useful AI question is measurable
Instead of asking how the organisation can use AI, ask where people are spending unnecessary time, where customers wait, where information is repeatedly rewritten, where decisions lack useful data or where a process produces inconsistent quality.
Those questions create a baseline. A baseline makes it possible to judge whether an AI use case actually improves anything afterwards.
Sometimes the answer is not AI
A broken process does not become a good process because a language model has been attached to it. Sometimes the most valuable intervention is clearer ownership, fewer steps, better information or a simple automation.
What good adoption looks like
Practical adoption means choosing a worthwhile use case, introducing it with appropriate human oversight and then measuring the result. Useful measures might include time returned, response speed, throughput, quality, confidence, adoption and customer experience.
Find the business problem first. Earn the right to talk about the technology afterwards.
Carl Kirby, KIST Consulting
