← Back to home
“AI executes at speed. It does not tell you whether the problem is worth solving — or whether you’ve defined it correctly.”
The core tension this channel is built around

What the current conversation is missing

Across forums, LinkedIn, and practitioner communities, the dominant question right now is: how do we build AI into our solutions? Which tools to use. How to automate workflows. How to generate outputs faster. How to embed AI into products and processes.

These are legitimate questions. But they are execution questions. They assume the problem is already defined — and defined correctly. In practice, that assumption fails more often than it should.

Teams have always been better at executing on a direction than questioning whether the direction is right. AI amplifies that tendency. It makes execution faster, more scalable, and more convincing. Which means a misdiagnosed problem can now scale further, faster, with less friction — and with more confidence — than ever before.

The risk is not that AI gives bad answers. The risk is that it gives confident, well-structured answers to the wrong question. And in organisations under pressure to show progress, well-structured confidence is often enough to move resources.

Why diagnosis is still a human and framework problem

Lean Six Sigma has a define phase for a reason. The first step in DMAIC is not to measure, analyse, or improve — it is to correctly state what problem you are actually trying to solve, for whom, under what conditions, and how you will know when it is solved.

Project management methodologies spend significant effort on scope definition, stakeholder analysis, and requirements clarification before any delivery work begins. PRINCE2 has a project initiation stage. The PMBoQ dedicates entire knowledge areas to scope and integration management.

These were not bureaucratic additions. They were added because experience showed, repeatedly, that work delivered on a misunderstood problem costs more to fix than it cost to deliver.

What AI does well

Execution at scale

Generating options, synthesising research, drafting outputs, analysing patterns across large data sets, and accelerating delivery once a direction is set.

What frameworks still do

Defining the right direction

Structured problem definition, stakeholder alignment, root cause analysis, scope boundary setting, and validating whether the problem is worth solving at all.

Where AI struggles alone

Diagnosis without context

AI does not know your organisational politics, your team's real constraints, your stakeholder's unstated concerns, or whether the stated problem is the actual problem.

Where frameworks still earn their keep

The messy front end

Voice of the customer, gemba walks, stakeholder mapping, root cause analysis — these are human and contextual activities that structure the problem before any solution begins.

How AI fits into the picture — as part of the solution, not instead of it

This is not an argument against AI. It is an argument for sequencing. When a problem is well-defined, AI is a powerful accelerant. It can help generate solution options faster, model scenarios, synthesise stakeholder inputs, and identify patterns that would take a human analyst far longer to surface.

The practitioners who will use AI most effectively are not the ones who start with AI. They are the ones who start with structured diagnosis — and then use AI to move faster once the direction is clear.

Why the six execution domains still matter in an AI-assisted world

Beyond Theory What Works is built around six domains that determine whether any project or improvement initiative actually delivers. These domains did not emerge from theory — they emerged from observing what distinguishes initiatives that land from those that stall or backfire.

AI does not eliminate the need for these domains. In some cases it makes them more important, because the speed of AI-assisted delivery can outpace the organisational readiness to absorb change.

Domain 01

Problem & scope clarity

Is the problem actually defined? Or has AI been asked to solve something that hasn't been correctly framed yet?

Domain 02

Stakeholder alignment

AI-generated outputs still need human buy-in. Stakeholder resistance does not disappear with faster delivery.

Domain 03

Risk & dependency management

AI amplifies delivery speed. Unmanaged risks and dependencies amplify at the same rate.

Domain 04

Measurement & baseline

AI pattern recognition is only as good as the data and baselines that exist before the work starts.

Domain 05

Change readiness

Speed of delivery through AI can outpace the organisation's capacity to absorb and adopt change.

Domain 06

Governance & accountability

Who is accountable when AI-assisted decisions go wrong? Governance clarity matters more, not less.

What this means for practitioners

If you lead projects, manage improvement programmes, or are responsible for operational change — your value is not in the tools you use. It is in your ability to correctly diagnose what is actually wrong, align people around a shared understanding of the problem, and govern delivery through to a measurable outcome.

AI makes the execution of a well-defined plan faster and more powerful. It does not make the diagnosis easier. It does not resolve stakeholder conflict. It does not tell you which problem to solve first when resources are constrained. These remain human, judgement-intensive activities — and they remain the critical path to whether any initiative succeeds.

The practitioners who will thrive are not AI sceptics or AI evangelists

They are the practitioners who understand which part of the problem-solving process they are in at any given moment — and which tools, frameworks, and disciplines are most appropriate for that phase.

Diagnosis first. AI inside a well-defined problem space. Execution excellence regardless of which tools are in the mix.

That sequence has not changed. Only the tools available in the middle phase have.

Not sure which approach fits your situation?

Beyond Theory Guide asks about your industry, role, project type, and constraints before recommending any framework, method, or approach — including where AI fits.

Try the advisor free →