The pitfall AI makes bigger
The question I get in almost every boardroom is whether AI can simply run the innovation process. Not assist, not suggest. Run. The answer is that AI can do the work in every phase of the CX+ Method, and in every phase it pushes you toward the exact mistake that phase was designed to prevent.
Five phases: Exploration, Problem, Solution, Proof of Concept, MVP. Each one exists because teams keep making the same error at that point. AI speeds all five up. It also makes the error easier to make and harder to see, because the output looks finished.
Exploration asks you to subtract
Exploration is the phase where you throw things away. You start with forty possible directions and you leave with three you actually believe in. That is subtraction, and it hurts.
AI adds. Ask it for market directions and you get twelve more, each with a plausible paragraph attached. Your list grows, your conviction shrinks, and the team spends three weeks reading instead of choosing. Use AI to gather the raw material, then close the laptop and pick with the people who carry the consequences.
The value of Exploration is in what you decide to ignore.
The Problem phase asks you to listen
Here the whole point is that you shut up and the customer talks. Interviews, observation, sitting in on service calls. Uncomfortable, slow, and the only way to find out what actually goes wrong.
AI is a talker. It will write you a problem statement before you have spoken to a single customer, and it will be articulate enough that nobody in the room objects. That statement is a summary of the internet, not of your customers.
What AI is genuinely good at here:
- transcribing and clustering fifty interviews you already conducted
- pulling recurring language out of service tickets and reviews
- spotting where your own documents contradict each other
The Solution phase asks you to choose one
Solution is where you let go of the concept you fell in love with. Most teams struggle with that. They keep three concepts alive because all three have a champion in the room.
Give those three to AI and you get three better ones. Sharper wording, cleaner value propositions, a slide deck each. The weak concept now looks strong, and the choice gets harder instead of easier. Judge concepts on customer evidence you collected yourself, and let AI polish only the one that survived.
Proof of Concept asks for realism
The PoC exists to find out what breaks. Cost per transaction, integration with your ordering system, what happens when a customer asks something nobody anticipated.
AI is the most cheerful colleague you will ever hire. It will tell you the approach is feasible, the architecture is sound and the timeline is achievable. It has no idea what your legacy system does on a Monday morning. Write down in advance what result would make you stop, and let a human check it against reality.
What you do on Monday
- Name the one decision each phase has to produce, and write it on the wall before AI touches anything.
- Let AI analyse what customers said. Keep the judgement about what customers feel with people who spoke to them.
- Book three customer conversations this week, before the team writes a single problem statement.
- Define the stop criterion for your current proof of concept in one sentence, with a number in it.
Running these five phases with AI in the room takes a way of working that holds under speed. CX+ The Method lays out the phases, the decisions and the checklists per phase, so your team knows what to hand over and what to keep. Bekijk de materialen