Chapter 8 — Four Possible Futures (plain-language version)

Building the four scenarios

Chapter 7 traced what AI does at each stage of an advisory career. This chapter carries that pattern into the future through four scenarios. They are not predictions. They are tests of what the argument of this thesis commits to if the world goes one way or another.

The setting is European infrastructure advisory, with its third box: the public and the future generations who live with infrastructure decisions without having signed any contract. But the two questions behind the scenarios apply to any advisory profession. Where does expertise live, in human practitioners or in AI systems? And who owns and governs those systems: the Anglo-American commercial platforms that draw knowledge out of the firms using them, or someone else?

A fifth exercise outside the grid, the Knowledge Commons, shows that ownership and formation, the long process by which a practitioner is made, are separate problems. Seven findings that hold in every future close the chapter.

Why thought-experiments, and not forecasts

Scenario planning 1 builds alternative futures so that an organisation can test its decisions against them. Philosophers use thought-experiments to see what a set of ideas commits them to under other conditions 2 . I use the scenarios both ways, and claim no one is likelier than the others.

Scenario one: Commodity Advisory

Expertise moves into AI systems owned by the commercial platforms. Advisory work becomes something any firm can reproduce with the same tools, and whatever made one firm different shrinks to a narrow band of senior judgment. The profession narrows; it does not disappear. Value flows up to the platform vendors. Junior formative work is absorbed, so the path by which seniors were made collapses, with consequences two decades on. The formal economic model of Acemoglu and colleagues describes the same path: a community’s shared knowledge vanishes even while each individual decision still looks good 3 .

Milan, 2024. An Italian infrastructure advisory firm loses three bids in a row to competitors offering AI-produced analysis, faster and cheaper. The partners are sure the winning work is thinner, and the clients do not seem to weigh that against the price. One partner titles a presentation: The market is pricing our distinctiveness at zero. 💡

The third box loses its footing. Advisers answer to the public through formed judgment, engagement with consequences and answerability over decades. When these are absorbed into generic tools, whatever answerability is left runs through systems tuned to the commercial platforms, not to the European public-interest tradition. Europe’s own way of doing public-interest work and each firm’s own voice both erode, and firms’ knowledge flows out to the vendors and back to their competitors. This scenario worries me most, and current conditions tilt towards it unless the institutions involved act.

Scenario two: the Digital Twin Economy

Expertise again moves into systems, but these are European and public, or built and owned by the firms themselves.

Western France, 2024. A port authority commissions a digital twin, a working digital model, of its logistics, built by a French consortium on French port data and kept within French jurisdiction. An advisory firm is brought in to advise on the rules for the twin and not on the port: who owns its outputs, who may question it, and what happens when its recommendations conflict with the port authority’s own judgment.

Here the European tradition and the firm’s voice both hold, and knowledge put into the systems stays recorded as belonging to whoever put it in. Junior work is still absorbed, so formation is altered, though it stays oriented to the European tradition. Senior work becomes supervision of the system. For the third box the outcome is uncertain: the systems know the public only as far as the public-interest tradition has been written into them, and that still has to be built in deliberately. The pieces the scenario needs, European public-interest AI, firm-owned AI and rules for both, exist only in early form.

Scenario three: Artisanal Advisory

Expertise stays in people, and the surrounding rules recognise the European public-interest tradition. AI is used, but bounded. Firms, professional bodies and public buyers deliberately protect the path through which juniors are formed.

Copenhagen, 2025. A Danish practice of twelve consultants has built its position on one commitment: all regulatory analysis is produced by a named senior consultant with experience in the client’s sector. It uses AI extensively for background research and not for drafting what the client reads. It charges a premium its larger competitors find implausible.

Here the third box is best served: advisers answer to the public through judgment formed the long way. The European tradition and the firm’s voice both hold, and the outflow of knowledge is contained, not stopped.

The risk lies in how the market prices this scenario. The HEC Alumni white paper expects some clients to pay more for advice without AI, as for a luxury handbag 4 . That version keeps formation open for a few firms and clients; the public-interest version keeps it open for the third box. Left to the market, the luxury version is the one already forming.

Scenario four: Disintermediated Clients

Expertise stays in people, but the systems belong to the commercial platforms. Sophisticated clients, such as utilities and regulators, build their own AI capability, and it absorbs the analysis they used to buy. Advisory shrinks to what in-house tools cannot reproduce: senior judgment.

Picture a utility that has brought the work in-house. Its advisory firm keeps four senior conversations a year, with no deliverables. Nine years of the firm’s knowledge now sit partly in the utility’s AI platform, which a vendor manages, and the board has not asked who owns the model. 💡

Extraction defines this scenario. The firm’s accumulated knowledge becomes training material for the client’s platform, on the vendor’s terms, so Europe’s tradition and the firm’s voice are both lost by another route. Junior and mid-career work shrink, and formation with them. For the third box the outcome is again uncertain. It depends on whether the client’s in-house capability can carry what advisers owe the public, and the assurance the client used to buy with an outside firm does not travel with the output. 💡 Of the four, this scenario most threatens the profession as a profession, and it is already emerging in some infrastructure sectors.

A supplementary thought-experiment: the Knowledge Commons

Imagine nineteen European organisations, public and private, that own and govern a shared AI platform trained on eleven years of their pooled project work. A junior analyst in Rotterdam, three years into her career, uses it daily and produces competent work. She has never built an analysis of her own that she did not abandon when the platform suggested something better, and she cannot say whether the judgment she exercises is hers.

The commons solves ownership. The members co-own what they put in, European public-interest practice remains the platform’s content, and its mandate can be directed at the public. It does not touch formation. The platform gives people fluency in the field’s language at scale. It cannot give contributory expertise, the ability to contribute to a field, which is formed only through long engagement with consequences 5 . Worse, because its outputs fit European practice so well, thin formation looks like solid competence across all nineteen organisations at once.

So ownership and formation are separate problems. The choices about formation that define Artisanal Advisory would be needed inside a commons as much as outside one.

Beneath the platform sits a second shared resource: the stock of formed practitioners. The profession’s own white paper forecasts that talent will be the scarce resource of the future 4 while recommending the automation of the junior tier throughout. Each firm that automates its own junior tier acts rationally, and each draws down the supply of seniors fifteen years from now, which every firm shares.

Seven findings that hold across all four scenarios

First, professional value turns on the difference between contributing to a field and sounding competent in it. AI reproduces the fluency at scale and cannot reproduce expertise formed through consequences 5 . The other six findings rest on this one.

Second, the question of formation cannot be avoided. In every future, how the profession makes its next experts is at stake, and the answer comes only from choices that firms, professional bodies and buyers make or fail to make.

Third, sovereignty is under pressure at more than one scale: Europe’s own way of doing public-interest work and each firm’s own voice, with different force in each scenario. Because sovereignty here means an identity at risk, the finding also says whose way of seeing survives: generic tools flatten it in Commodity Advisory, firm-owned systems carry it in the Digital Twin Economy, Artisanal Advisory keeps it in the work and in each practitioner’s own signature, and it thins with the firms themselves when clients take the work in-house.

Fourth, the third box persists. The people who bear the consequences without holding a contract, including the future generations whose conditions of life are being shaped now, do not go away in any scenario. Only who is made to answer for them changes.

Fifth, legitimacy has to be spelled out. Credentials and seniority no longer carry it on their own, and a firm must say what was done, who did it, what the AI did and what the practitioner contributed. So far firms tell that story to financial markets, as a story of control 6 . That answers for the firm’s viability, not for what the work owes the publics who live with its results.

Sixth, extraction happens in all four. Knowledge flows out of advisory practice in each, and the live question is on what terms of ownership.

Seventh, the Applied stage, the middle years of a career where analysis connects early formation to senior authority, is safe in no scenario. 💡

The ground the scenarios leave

The scenarios differ in how the seven findings play out; the findings hold in all of them. Chapter 9’s recommendations answer to that shared ground, and go to the people and institutions whose decisions shape which future arrives.