What MCG Finds in the First Thirty Days that Changes Everything

Every organisation arrives at MCG believing it is further along than it is. The gap between what leadership thinks is ready and what the data actually shows is the finding that changes everything. Marques Consulting Group on what Phase One assessments consistently reveal, and why the diagnostic has to come before anything else.

THE DIAGNOSTIC

Michelle Margaret Marques

9/7/20264 min read

Senior female executive leading a corporate training workshop for diverse employees using a whiteboard and data charts.
Senior female executive leading a corporate training workshop for diverse employees using a whiteboard and data charts.

Every organisation that comes to MCG for an AI engagement arrives with a version of the same belief. They are further along than they are.

Not because they are being dishonest. Because the gap between what an organisation believes about its own readiness and what the data actually shows is almost always invisible from the inside. The people closest to the work have built their understanding of it through the work itself, and that understanding is accurate as far as it goes. It does not go far enough.

What the first thirty days of a Phase One assessment consistently reveals is not that the organisation is unprepared. It is that the preparation that has happened is almost entirely at the wrong layer.

What organisations think they are ready for

By the time an organisation engages MCG, it has usually done something. It has run pilots. It has evaluated platforms. It has identified use cases. It has, in many cases, already spent significant budget on tools that are technically operational but not producing the return the business case promised.

The readiness narrative that arrives with the engagement is almost always structured around these visible markers. We have the technology. We have identified where we want to use it. We have leadership buy-in. We are ready to build.

What that narrative describes is readiness at the capability layer. The technology exists. The appetite exists. The intention exists. What it does not describe, and what it almost never accounts for, is the layer underneath capability where the actual build has to happen.

What the data actually shows

The consistent finding across Phase One assessments is not a single gap. It is a cluster of gaps that exist in a specific relationship to each other, and that relationship is what makes them difficult to see from the inside.

The workflow that leadership believes is ready for AI implementation is almost never the workflow that actually runs. There is the version in the process document, the version that was designed, the version that makes sense on a slide. And then there is the version that actually operates day to day, which has adapted around the constraints, the workarounds, the informal decisions, and the institutional knowledge that never made it into any document. Building an AI system onto the documented version of the workflow produces a system that is technically correct and operationally irrelevant.

The data that the organisation believes is ready to support the use case is almost never in the state required to carry it. Not because the organisation has been careless. Because data accumulates in the shape of the processes that created it, and those processes were not designed with AI implementation in mind. The gaps, inconsistencies, and structural issues in the data are not visible until they are specifically looked for, at the level of granularity the system will actually require.

The accountability structures that leadership assumes will govern the system do not yet exist at the operational level. There is awareness that governance matters. There is, in many cases, a governance framework at the policy level. What does not exist is the named accountability at the workflow level, the specific human decision points, the defined escalation paths, the documented permission boundaries that determine what the system is and is not authorised to do before it does anything.

And the organisational readiness for what the system will change, the team-level understanding of how the work will shift, who will review what, what the new workflow looks like for the people inside it, is almost always at a lower level of preparation than leadership believes it to be. Not because the teams are resistant. Because nobody has yet had the specific conversations that make the change real rather than abstract.

Why this is always a surprise

The gap between what an organisation believes about its readiness and what the assessment reveals is not a failure of intelligence or due diligence. It is a structural feature of how organisations understand themselves.

Leadership sees the organisation from the altitude at which leadership operates. That altitude produces an accurate picture of strategy, capability, and direction. It does not produce an accurate picture of how work actually moves at the operational level, where the data actually lives and in what state, where the informal decision-making actually happens, and where the assumptions that are being carried into the AI initiative have not yet been tested against operational reality.

The assessment goes to the level where the build will actually happen. And at that level, the picture is almost always materially different from the one that arrived with the engagement.

What changes when the assessment comes first

The organisations that build AI systems without a rigorous Phase One assessment build on the visible version of their readiness. The documented workflow. The assumed data quality. The policy-level governance. The leadership-level understanding of operational reality. What they build reflects those assumptions, and when the system meets the actual workflow, the actual data, and the actual decision-making environment, the gap between the two is what determines whether the system holds or fails.

The organisations that begin with the assessment build on what is actually there. The workflow as it operates. The data as it exists. The governance as it needs to be designed, not as it is assumed to be. The team readiness at the level where the work happens.

The assessment does not add time to the project. It removes the time that would otherwise be spent rebuilding something that was built on the wrong foundation.

What MCG finds in the first thirty days does not change what needs to be built. It changes what the build is actually for, what it has to account for, and what the system will be capable of holding once it is live.

That is why the assessment has to come first. Not as a preliminary step before the real work begins. As the first and most consequential piece of the real work itself.