Chapter 4 — How Expertise Is Actually Built (plain-language version)

What this chapter is for

Nobody arrives in European infrastructure advisory already expert. Expertise is made, by a system working at three scales that produce one another: the individual consultant, the firm as a collective that holds knowledge and passes it on, and the domain, meaning infrastructure advisory as a tradition built around the public interest. The individual is formed by the firm and the domain, the firm by the people it trains and the domain it works in, and the domain by the firms that work within it.

The chapter is descriptive. Chapters 5 and 7 ask what AI is doing to this system; here I set it out as it has historically operated. 💡

The individual: how a consultant comes into being

A consultant is not a finished professional type but someone who keeps becoming a particular kind of practitioner over a career, a process that Stiegler, a French philosopher, called individuation 1 2 . Ibarra, who studied young consultants and bankers, watched that becoming at close range: juniors observe role models, try out provisional versions of themselves, and judge the results against their own standards and others’ feedback 3 . The mechanism is the same in every advisory field; only the content changes. It has three parts.

The first is formal training. It supplies tools and vocabulary, and it is necessary but not sufficient. 💡

The second is work that has consequences: building a tariff model from a blank page and learning from the corrections, drafting memos that fail. This builds what Collins, a sociologist of expertise, calls contributory expertise: the ability to contribute to a practice, as opposed to talking about it fluently 4 .

The third is senior judgment, watched and absorbed: watching seniors, Ibarra’s juniors built “a store of tacit knowledge, attitudes, routines, impression management techniques” 3 . Much of it is what Polanyi, a philosopher, called tacit knowledge: picked up by working alongside someone, and not always possible to put into words 5 .

Brussels, 2020. A team has a sound case for a formal challenge to the Belgian energy regulator. The senior partner, who has worked with that regulator for two decades, says “not now” and cannot fully explain why. Eighteen months later the regulator opens a sector-wide investigation that hits the client’s revenue. He had read something that was in no document. 💡

The Chartered Management Consultant (ChMC) framework sets this formation in three career stages: Foundation, where formative work is concentrated; Applied, where autonomy grows; and Chartered, where judgment becomes the main contribution 6 .

Formative work also gives disorder: the wrong structure, the draft that does not land. The profession’s own HEC Alumni white paper calls this structuring chaos (“chaos structurant” in French) and observes that juniors handed polished answers by an AI assistant no longer pass through it 7 . It is doing the work badly, under correction, until the corrections become one’s own. 💡

The firm: a collective that manages knowledge

To describe how firms hold knowledge I use the SECI model of Nonaka and Takeuchi 8 9 . It follows knowledge as it moves between people’s heads and written form: shared through joint work, put into words, recombined, and absorbed back into practice, in a continuous loop.

A mid-sized London firm has a concept it calls “the regulatory clock”: an informal way of reading where a regulatory process is in its cycle. It is in no playbook. An analyst who joins is confused for three months by its absence from any document, uses it by instinct by month six, and by year three teaches it to a newcomer. 💡

I read this loop as producing identity, not only productivity. Through it a firm reproduces its vocabulary, its apprenticeship, its memory of past cases and its methods: the things that make it a particular kind of firm.

In large advisory firms, the asset is this knowledge architecture and the pipeline that turns juniors into seniors, and both are exposed when AI takes over formative work. The pyramid, in which the billing of many juniors paid for their formation, needs junior work to bill, and many firms are already hiring fewer juniors with no real plan for what that does to the pipeline 10 . 💡 The profession’s proposed replacement is an obelisk: the same summit on a much narrower base 7 . It answers a question about margin. An obelisk has seniors and no account of where they come from. 💡

In a boutique, the value sits in the personal judgment of one or a few senior people. The risk falls on the person: whether an adviser using the same AI tools as everyone else can keep a voice of their own.

Neither kind of firm keeps its adaptations to itself. Killewald and Haskamp’s study of two European consultancies finds that with the tools in use “practice loops are shortened or skipped” 11 while value moves towards collected experience 11 . The field now prizes the very thing the shortened practice stops producing.

The domain: what infrastructure advisory actually asks of people

The first two scales look alike across advisory fields. The domain is specific: the mechanism is general, and the evidence of this thesis comes from European infrastructure.

Four features define it. It is a family of sectors, from energy and water to transport and defence, that differ in how heavily they are regulated and how far ahead they must look. It is held together by a public-interest tradition. The third box is this thesis’s name for the people that tradition answers to: the public and the future generations who live with infrastructure decisions without being party to any contract, and to whom advisers owe an answer. The domain joins technical depth to political judgment. And its knowledge is contested: the same project can be judged by cost, fairness, climate or strategic autonomy.

London, 2019. A firm produces an economic assessment of a third airport runway. Before a parliamentary committee, the senior partner is asked whether the analysis assumes compatibility with the national carbon budgets, and answers that this is a separate policy question. He is technically correct. He is also taking a side on whether advisers answer for the consequences of their framings for the publics affected, or only for the correctness of their models.

AI arrives in a domain already reshaped over three decades by deregulation, privatisation and climate policy. 💡

How the three scales make one another

The Hague. A Dutch practice founded in 1998 works almost only on water, and has come to treat a water asset as a service to be delivered first and a cost to be recovered second. A partner arrives in 2019 from energy regulation, trained in rate-of-return thinking. Within eighteen months she frames water problems in the language of service delivery, though nobody told her to: the clients ask in that language.

The individual’s professional identity is produced by the work the firm makes available, by its supervision and by the problems the domain presents. The firm’s identity is reproduced through the people it forms and shaped by its domain. The domain is what its firms have been making of it. So a change at any one scale spreads to the other two: change how consultants are formed and you change the firm, and then the domain.

The whole system in one view

European infrastructure advisory builds its expertise at three scales that make one another. Individuals are formed by training, consequential work and absorbed senior judgment. Firms turn unwritten knowledge into shared assets, while apprenticeship supplies what those assets cannot hold. The domain ties both to something larger than any firm: the European regulatory tradition and the third box, the public and future generations whose conditions of life the work shapes.

Chapter 5 asks what AI is doing to this system, and whether firms’ knowledge is being drawn out into the vendors’ systems 💡 .