Discover where AI belongs. Decide what deserves to move forward.
Right Problem Method™ discovers valuable opportunities in real operational work, challenges whether AI is actually the right response, and defines what is worth pursuing.
Where does AI actually belong?
Created by António Rocha · Head of AI and Project Manager at Ambidata · Professor of Computer Engineering at ISEP · 25+ years across software engineering, requirements and delivery
The right solution starts with the right problem.
Start with the work. Understand what matters. Move through four structured phases to decide what deserves to move forward.
Most AI conversations start too late.
Many organisations now generate AI ideas faster than they can evaluate them. The result is a crowded pipeline with no clear view of which problems are worth solving, which ones require AI and which initiatives should be funded.
The problem is rarely a shortage of AI ideas. In many organisations, there are already more proposals than the business can properly evaluate, fund and implement.
The question should not begin with:
“What can we do with AI?”
It should begin with:
“What problem is worth solving, and where does AI actually belong?”
Start with the problem, not with the assumption that the answer must be AI.
AI should not be the starting assumption.
Right Problem does not begin with a list of technologies or with a workshop whose purpose is to invent AI use cases. It begins with the organisation, its readiness, its real processes and its operational problems.
The purpose is to establish where AI belongs and which opportunities merit further investment.
Diagnose. Find. Interrogate. Target.
The method uses four clear stages. It starts with organisational readiness and operational problems, then filters opportunities and defines the strongest candidates.
Can the organisation support meaningful change?
Set the sponsor, scope and success criteria. Then test whether the organisation can absorb the change and identify the conditions that could affect delivery.
Sponsorship · decision discipline · process clarity · change capacity · data accessibility
Where is the real operational problem?
Look at how work is actually done instead of asking teams to invent AI ideas. Identify friction, exceptions and measurable business problems.
Observe actual work · expose exceptions · anchor value · map the real process
Does this problem really justify AI?
Challenge every candidate before it can become a recommendation. Does it create material value? Is it really AI? Build or buy? Is it sufficiently defined? Is the organisation ready? Are there relevant risks or regulatory concerns?
Value · AI fit · build or buy · definition · readiness · risk
What must be defined before it can move forward?
For each initiative that survives, produce a specification that supports estimation, design and an investment decision.
Business value · requirements · production integration · governance · adoption · sequencing