The Consultant's Workbench

For three decades, consultants sold one thing: diagnosis packaged for presentation. Frameworks, matrices, recommendations. Value lay in the quality of the analysis and the elegance of delivery. Execution was left to the client, their IT teams, or a third-party integrator. Between the thinker and the builder, the line was clear — and it structured the entire profession.

That line is shifting. Not everywhere, not for everything, but along a precise zone that needs naming before going further.

Where the line moves — and where it doesn't

What is changing is not a shift in professional ideology — it is a shift in economics. Until recently, building a bespoke tool cost enough that the calculation almost always tilted toward off-the-shelf SaaS. Better to rent a tool used by ten thousand companies than to fund a development team for six months to address a specific need.

With generative models, the cost of creating and iterating a custom tool has dropped sharply. What took six months now takes two weeks. What required a team now requires one person who can both read a business need and hold their own in a code editor.

One must immediately bound what this does not mean. The zone under pressure is not the full SaaS stack of an SME. It is the intermediate layer: reporting and consolidation, business workflows, specialist assistants, interfaces between existing systems, local automations. This layer is precisely where generic standards chafe most against real work, and where every company ends up paying for a generic subscription to do roughly what it actually needed. That is where the consultant's workbench becomes credible. Outside this zone — for ERP, regulatory accounting, messaging, the critical tools at the core of the information system — SaaS remains structurally superior, and for good reasons.

What SaaS carries that AI does not reduce

SaaS does not win simply because it costs less to build. It wins because it carries a burden that gets forgotten when you look only at the creation cost. Corrective and evolutionary maintenance, security updates, regulatory compliance, user support, continuity in the event of failure, integrations maintained across the full range of third-party tools. This burden is real, it is continuous, and it has a cost.

Generative AI dramatically reduces the cost of creating and iterating a tool. It does not, by itself, reduce the cost of responsibility that follows deployment. That responsibility still falls to someone — now, in the model I am describing, to the consultant who delivered the tool. This is what distinguishes a prototype demonstrated in a meeting from a tool that runs for three years without waking its author at two in the morning. The new-model consultant does not sell a prototype. They sell a responsibility held over time, within a perimeter they have chosen because they can genuinely hold it.

Immersion replaces the engagement

The practice begins with immersion. The consultant moves into the client's environment, observes the flows, watches what employees do when they work. This observation is no longer limited to interviews and workshops. It can draw on tools that measure, at the level of individual workstations, the tasks that take time and repeat themselves. This map of activity, crossed with an understanding of the business, reveals the gestures that deserve a tool — and those that do not.

This kind of instrumentation calls for a discipline that needs to be named clearly, because the risk of drift is real. The observation is consented to, announced to the teams before it is deployed; it aggregates rather than names individuals, and its purpose is bounded to identifying tasks to automate. It is not used to evaluate people. It stops when the mapping work is done. Instrumentation that moved outside this frame would be surveillance — and the trust contract with the teams, which is precisely what makes immersion possible, would be broken.

What the consultant then delivers is no longer a document describing what should be built, but a working tool. That tool replaces, where relevant, a SaaS subscription that weighed on cash flow without serving real work. It is sized to the client's actual flows, not the generic flows that SaaS imposes on everyone. Bespoke in practice replaces the generic slide machine. At the scale of an SME budget, the sum of subscriptions in the intermediate layer regularly amounts to several percent of turnover; the recoverable share is immediate and quantifiable.

The distinction that must be held

The deeper benefit is not in the invoice — it is in the time returned to employees: time they can invest in what justifies their presence in the company, rather than in manually steering a piece of software that doesn't fit. That effect, in my view, is the one that genuinely changes the nature of the gain; it shifts the argument from an accounting register to a professional one.

The market conflates two things that must be distinguished. AI transformation of a business is not the same thing as training employees to use AI. You learn to use an assistant by using it, and generalist awareness seminars — those that repeat what everyone has already read — add little. Targeted acculturation, by contrast, conducted around the tools the company is actually deploying, for the real work of the teams, retains its full value; it is even what determines whether the delivered tool will actually be used. The issue is not coaching — it is coaching with no grounding. Transformation demands something else: a rare competence capable of understanding both a business value chain and the technical tool chain that can sustain it. That is the competence the new-model consultant must offer, and it is the one that cannot be pooled in an online session.

The profession does not disappear — the job description transforms

The consultant does not disappear any more than most professions do. But the job description changes radically. What was sold — producing frameworks, delivering analysis — loses relative value. What is sold now — immersion, field diagnosis, delivery of tools sized to real flows, and responsibility for maintaining them — demands competences that were not, until yesterday, part of the profile. Understanding an organisation and managing a technical tool chain have become the same qualification. A share of consulting will remain in commentary and lose ground to those who deliver the tool alongside the analysis.

For territories that are not strongholds of consulting, this reconfiguration is a window. The new consultant needs to be close to the client, to know their real constraints, to iterate quickly. These conditions favour local presence over metropolitan centralisation — and when the tool built rests on a sovereign inference infrastructure, the proximity advantage becomes a sovereignty advantage.

The shift is not incidental. It redefines what a consultant sells: no longer advice about what should be done, but the tool that does it, and the responsibility to maintain it. The workbench replaces the firm. And what comes off the workbench holds together from the moment of delivery — and the morning after.