Turn AI ambition into decisions you can defend.

No. The method starts with the problem and the work, not with a technology. It can be used when your organisation is considering AI, when a process is creating measurable friction, or when an important investment decision needs stronger evidence.

AI is considered during the method only when it is a credible part of the answer. If another approach is more appropriate, the diagnostic should make that clear.

A Right Problem Diagnostic is a structured way to understand an organisational context, examine how work is actually done, test the value of potential opportunities and define what deserves attention next.

The diagnostic follows four stages: Diagnose, Find, Interrogate and Target. Each stage answers a different question and adds evidence to the decision. The purpose is not to produce a long list of ideas. It is to establish which opportunities should proceed, which should stop and what needs to be understood before making a responsible commitment.

A diagnostic makes sense when your organisation wants to discover where meaningful opportunities exist, based on how work is actually done. It is especially useful when teams are experiencing friction, repeated exceptions, avoidable manual effort, delays or lost capacity, but the strongest opportunity has not yet been clearly defined.

It can also help when you already have AI proposals that need to be tested, when a known problem still lacks a convincing solution, or when a significant investment is being considered. The starting point is not a list of technologies. It is the work, the problem and the evidence that shows where attention may be worthwhile.

No. The Right Problem Method creates direction, not dependency.

The diagnostic establishes what the opportunity is, why it matters, what kind of response it needs and what must be defined before work can move forward. Your organisation can then continue internally, work with an existing partner or select a new delivery partner with a clearer brief and better decision criteria.

The work begins by establishing the context, sponsor, scope and conditions for a useful diagnosis. We then examine how the work is actually performed with the people closest to it.

Potential opportunities are challenged against consistent questions. We look at the value of the problem, the evidence behind it, the direction of the solution and the role of AI. We develop the strongest opportunities only as far as the evidence justifies and make every direction explicit.

The main outcome is a Right Problem Brief. It brings together the relevant evidence, findings, rationale and decision outcomes needed to support the next investment decision.

For the opportunity that moves forward, the diagnostic also produces a final Specification Document. It defines the opportunity clearly enough to support estimation, solution design and a responsible decision about what should happen next. The specification goes only as far as the evidence justifies. It is not a build project or a substitute for detailed implementation design.

Depending on the opportunity, the wider set of outputs may include organisational readiness, process findings, prioritised candidates, solution direction, explicit decision codes and the information needed for estimation or further design. Not every opportunity needs the same level of definition.

Yes. That is a valid and sometimes valuable conclusion.

The method does not assume that every worthwhile problem needs an AI solution. An opportunity may receive a NOT-AI direction when the problem matters but another response is more appropriate. A no to AI is not necessarily a no to the problem.

Each opportunity receives an explicit direction rather than disappearing into an unexamined backlog.

GO means the opportunity is strong enough to continue towards a defined next step. NO-GO means it should stop under the current conditions. NOT-AI means the problem matters, but AI is not the right answer. NOT-BUILD means the need is real, but buying, adapting or changing an existing solution makes more sense than building something new. NOT-VALUE means the expected value does not justify further investment. NOT-DEFINED means there is not yet enough evidence to make a responsible decision.

The purpose of these outcomes is not to score ideas for their own sake. It is to make the reasoning behind each direction visible.

The diagnostic requires an executive sponsor with enough authority to support the work and act on the findings. It also needs access to the owners of the relevant processes and to people who perform the work in practice.

Depending on the context, technology, data, compliance or other specialists may also need to contribute. The important point is that the diagnostic cannot be based only on strategy documents or assumptions. It must include the people and evidence closest to the work.

A typical Right Problem Diagnostic takes around four weeks. The exact timetable depends on the agreed scope, the number of processes, the availability of participants and access to the information needed for the work.

The engagement is normally organised around a focused scope, usually two related processes within one business function. The schedule and working arrangements are agreed at the start, including the participants, success criteria and access required. A narrow and realistic scope usually produces a more useful decision than a broad review that cannot examine the work properly.

Confidentiality and responsible access to information are discussed before the diagnostic begins. The scope, participants and information required are agreed in advance, and the work is limited to what is necessary to understand the opportunity and support the decision.

The method may also be developed through learning from real engagements. Any use of a client case for research, teaching, publication or the future Right Problem book requires a separate and transparent conversation. The client’s name and the names of its people are not used without specific permission, and any proposed publication is shared for review and approval before it is released.