The lost pitch
There is a small, almost incidental scene in every consultancy and every agency that captures the subject of this essay rather neatly. A pitch team stands in the meeting room on a Friday afternoon, closes the laptops and quietly congratulates itself on a thoroughly considered performance. The pitch is lost. These things happen. The client's reasoning was diplomatic, the internal debrief brief, the drinks afterwards a little longer. On Monday the week begins with three new enquiries, and the pitch becomes what pitches usually become in most houses. A folder on a drive, a slide in a deck library, a vague impression among the four or five people who, over the following two years, will find themselves in pitch situations adjacent to the same theme.
Eighteen months later a different team stands in a different meeting room in front of a similar industry, a related question and a client who uses almost the same words as the client from before. The new team does not know the earlier pitch. It does not know it exists. Even if it found the deck, it would not know that the argumentative arc on slide nineteen had been thrown out by the senior partner, revived in a second iteration and finally sharpened with a single line the team had only landed an hour before the meeting. The new team builds its pitch from scratch, politely, competently and, two days later, under pressure. It either wins or loses. In neither case does the firm walk away with anything it did not already have.
This scene is the prototype of the topic in front of us. In consultancies and agencies, the most valuable knowledge is generated in situations that are never recognised as knowledge production. Pitches, workshops, strategy sessions, design critiques, project reviews and handovers are the actual moments of value creation. Most houses treat them as project by-products and close them on the day the project closes. What survives is a body of decks, documents and notes that is technically available and practically unread, the moment the next mandate has captured the firm's attention.
Why consultancies and agencies are particularly exposed
An earlier essay in this series examined the same mechanism in law firms, tax advisory and audit practices. The structural resemblance is obvious. There too, intellectual work is produced under time pressure and filed in places that are seldom revisited. The differences, however, are significant, and they matter for any honest reading of the consulting and agency business.
Consultancies and agencies work in much shorter cycles. Where an audit relationship runs for five, ten or twenty years, a consulting project typically lasts three to nine months and an agency mandate follows a campaign logic of weeks to a year. Team compositions shift accordingly. In large consultancies a consultant cycles through six to ten projects within three years; in a mid-sized agency the figure is many times that in briefings alone. Knowledge that is not designed from the outset as shared knowledge inevitably becomes the private property of individual practitioners under these conditions.
The professions also lack a regulatory obligation to document. Audit firms document because they must. Consultancies and agencies document because a senior has, on a good day, asked them to, or because the firm's methodology demands it. Discipline therefore rests on leadership behaviour and house culture rather than on any external rule. The result is an unusually wide spread of documentation quality, from the impeccably maintained practice group to the project whose decisive insight lives in a single Slack thread.
Finally, knowledge in consultancies and agencies is narrative by nature. A strategic recommendation is, at heart, a story about a market, a competitive set and a direction of travel. A creative concept is, at heart, a story about a brand, an audience and a promise. Such stories live on tone, sequence, omission and punchline. They can be forced into a PDF. They regularly lose, in that transition, what made them more than the sum of their slides. Anyone reading an old strategy without having watched the original senior present it tends to see only the slides.
Pitching and business development as the most expensive knowledge loss in the house
In most consultancies and agencies, business development is the most expensive internal process that nobody really measures. Industry observers such as Source Global Research and national bodies including the BDU in Germany place average pitch win rates for management consultancies somewhere between twenty and forty per cent, depending on segment and seniority. In classical agency work, where bodies such as the GWA and the ADC publish regular reports, the rates are often lower, with the larger creative pitches frequently fielded against four or five competing houses and rewarding precisely one.
On the cost side, the figures are unloved in polite conversation. A mid-sized pitch in a strategy consultancy ties up seven to fifteen person-days of senior and junior capacity. A larger creative pitch in an agency, including production and research, can quickly run into the low six figures of internal cost. A house of fifty consultants or eighty creatives produces a three-digit number of pitches and proposals a year. Most of that work flows back into firms that do not consolidate it in any systematic way.
The real problem is not the individual lost pitch. It is the structural inability of most houses to derive, from twelve months of pitch history, a coherent picture of their own argumentative patterns. Which arcs work in which industries, which cases get requested in which sequence, which pricing questions come up when, which objections reliably surface in which constellations. All of this information sits inside the house. It lies in the decks, in the client follow-ups, in the email threads, and in the memory of the people involved. None of it sits anywhere that would allow the next pitch team to retrieve it within an hour.
A concrete example gives the dimensions a shape. A strategy consultancy with a focus on industrial and mobility clients pitches forty-seven mandates in a single year. Fourteen are won, thirty-three lost. The internal debrief follows a slim standard template that records the perceived effect of the pitch on a five-point scale and offers space for three free sentences. Anyone preparing a similar pitch the following year has access to the templates, yet has no way of knowing which argumentative sequence actually carried the room. That information sits in the deck of the second iteration, in the oral debrief with the client and in the recollection of the two partners present. Anyone hoping to brief the next team competently needs all three sources. Anyone unable to retrieve them starts from scratch.
In agencies, the pattern repeats with an added twist. Creative work depends on not repeating itself. A concept that convinced a brand three years ago cannot simply be recycled for the next brand. What can be recycled, however, is the methodological path that led to the concept. The choice of trend sources, the sequence of workshops, the way insights were condensed, the logic by which visualisations were arrived at. This methodological repertoire is the actual asset of a creative agency. In practice it lives in the heads of a handful of creative directors and strategy leads. When they leave, the repertoire leaves with them, and the next pitch team begins by improvising a fresh variation on a method the house has long since mastered.
Project methodology and the frameworks that nobody ever uses twice the same way
Consultancies and agencies live off methods. A growth strategy follows one logic, a brand transformation follows another, an operational excellence initiative follows a third. Most houses have spent considerable energy over the past two decades codifying these approaches as methods, often under brand names designed to make the house externally recognisable. On the website the methods sit in a tidy row. In projects, they are rarely applied in quite that form.
The reason is not negligence. It lies in the nature of the work. A framework that worked for a pharma engagement needs adjustment, in several specific places, for an energy mandate. These adjustments are the actual intellectual contribution of consulting life. They emerge in workshops, in late-night deck sessions and in the corrections of experienced partners. They become productive in the project and are filed, at project close, in a folder whose very existence is forgotten by everyone except the immediate participants within six months.
The result is a house that officially works with thirty methods and, in practice, with three hundred method variants, none of which are systematically captured. The fifth market analysis for an industrial mid-cap client typically starts again from the beginning, despite the firm holding at least four comparable analyses whose methodological adjustments and failed attempts would substantially accelerate the work. Those adjustments, however, sit in the decks of the respective projects, often in the comment trail of a third iteration, and often in the notes of the lead consultant, who has since left the firm.
In creative agencies the same pattern surfaces in its own form. A design process, a concept architecture, a workshop dramaturgy is reinvented for every larger mandate, even though the house has tested most of the building blocks in earlier work. Those building blocks live on the boards of past design teams, in the Miro workspaces of bygone projects and in the personal archives of senior creatives. A junior strategist planning her first workshop series writes it up from textbook knowledge, because the actually available house knowledge is not accessible to her.
Research and industry intelligence as the invisible library
In every consultancy and every agency, a significant share of value creation happens in research and synthesis phases. Market analyses, competitive landscapes, trend distillations, customer studies, interview banks, data work. Such work is produced inside the project context, delivered to the client, and then handed over to the internal filing system. In theory this gives the firm a house library on which every new engagement could build. In practice, nobody uses this library, because as a search surface it does not function.
A simple example illustrates the point. A consultancy has, over the past five years, produced eleven market studies on the European logistics sector, some for logistics providers, some for shipping industries, some for investors. Each study contains its own picture of the market, with its own data sources, assumptions and conclusions. As the twelfth study begins, the lead consultant would ideally start with a thematically ordered overview of the eleven predecessors. She would see which data sources had proven robust in which contexts, which assumptions had quietly cracked over time, which arguments had carried the client conversation, and which had not.
That overview exists in no ordinary document management system. It also fails to materialise in the classical knowledge management tools the industry has tried for twenty years, most of which have foundered on the same requirement, namely that the author of a study must anticipate the way in which it will later be retrieved. This act of anticipation is precisely what makes manual tagging in advisory work so unreliable.
In agencies, research work is often even more ephemeral. Trend observations, cultural insights, audience studies, social listening and competitive analyses are assembled for a pitch or a concept, presented and then never systematically consolidated. An agency that produces forty pitch and concept trend distillations a year is sitting, at year-end, on forty parallel pictures of the same cultural moment, none of which have been put into relation with one another. That is not merely inefficient. It denies the agency the chance to develop, out of its own body of observation, a differentiated industry view it could carry into the market.
Senior knowledge and succession in the quiet hour of truth
In consultancies and agencies, a remarkably large share of house knowledge lives in the heads of a handful of senior profiles. In strategy consultancies it is the partners and principals whose judgement carries a project. In creative agencies it is the executive creative directors, strategy leads and account directors whose instinct decides, in roughly five seconds, whether a concept lands for a given brand. That knowledge is, quite literally, irreplaceable, because the firm has no means of reproducing it once its carrier has handed in the laptop.
The demographic reality makes the question more urgent. The annual reports of Source Global Research on the European consulting market, and the regular industry surveys around the agency landscape, show that attrition has remained at elevated levels for several years. In large consultancies the average tenure at consultant level sits at three to four years and at senior level at six to eight. In agencies, junior tenure is often below two. Every departure, in the light of the preceding pages, is a knowledge outflow whose magnitude the firm rarely quantifies.
In most houses this outflow is addressed by a handover routine that begins, on a generous interpretation, two hours before the departing person's last day. The product of this routine is a brisk list of live mandates, a status note that nobody will reread and a vague suggestion of the relevant contacts. What is not handed over is precisely the layer the firm ought to care about. The unspoken rules of engagement with three key clients, the argumentative arcs that have carried in past pitches, the methodological adjustments that have rescued a difficult engagement. That layer walks out with the person and waves politely on the way.
The question of how a firm secures this layer for the next generation remains, in most consultancies and agencies, unanswered. It is usually deferred to a mentoring culture and to well-run project reviews. Both are useful and neither is sufficient. Mentoring carries only as long as the mentor remains. Project reviews carry only as long as the participants' memory window holds. Both collapse the moment the people involved leave the house or simply move on to the next project.
The unofficial AI reality in consultancies and agencies
In nearly every consultancy and agency, the individual response to the problem described above has long since arrived. It goes by the name of public AI. Junior consultants draft the first cut of a market analysis on a freely available language model. Strategists have their trend distillations pre-formulated and then rework them with a creative flourish. Account directors run client emails through public tools before sending. Senior partners sketch pitch structures on the same tools, often with rather more discretion than their juniors and no less regularity.
From the individual perspective this practice is entirely rational. It delivers visible acceleration, lowers the entry barrier into unfamiliar subjects and tends to lift the linguistic quality of the end product. From the firm's perspective it is problematic, and in a manner that differs from the legal and audit professions. In law firms and audit houses the central concern is professional secrecy. In consultancies and agencies it is, additionally, the outsourcing of house thinking.
Every interaction a consultant has with a public model is an interaction in which the model is enriched with the consultant's knowledge, without the firm participating in that enrichment. The body of experience accumulating in such interactions over months and years belongs, in any condensed form, neither fully to the consultant nor at all to the firm. It sits with a third party whose commercial logic gives that aggregation a particular meaning. A firm that does not pull this body back into an environment of its own is quietly handing over a meaningful share of its intellectual substance to an external actor that also serves its competitors with the same substance.
A contractual angle is rarely discussed openly in the industry. Client contracts in consulting and in agency work often contain confidentiality provisions that make the processing of project content in third-party systems legally non-trivial. A practice in which individual employees routinely paste extracts of project material into public models sits in a grey zone whose cost becomes visible the moment a larger client requests formal disclosure of the tooling landscape in use. That request will become a standard component of nearly every serious RFP within the next two to three years.
What a Private AI Knowledge System actually delivers in this industry
In this situation, a Private AI Knowledge System is not one more tool to be wedged into an already overburdened toolscape. It is the organisationally responsible version of what the staff is doing already. It moves the behaviour that today happens in public models into an environment owned by the firm, one that knows its own material and takes its confidentiality obligations seriously.
For a consultancy this means, concretely, that the firm's pitch and project archive becomes searchable along thematic affinity rather than folder hierarchy. A new pitch team begins preparation with a thematically ordered overview of every pitch over recent years that has moved in a similar direction, including the argumentative arcs that have carried, the objections that have regularly surfaced and the lead figures who can be approached for follow-up. The creative work of the new team is not replaced. It begins at a higher level and in a shorter time.
For an agency it means that the methodological repertoire and the research stock of recent years actually become retrievable. A strategy lead designing a workshop series for a new brand sees, in a moment, which workshop dramaturgies have lately been tested in comparable mandates, which insight distillations have worked and which visualisation logics the client in question has tended to welcome. A creative director seeking to place a new concept gains an overview of what the agency has previously thought about the same cultural field. The finished concept is not standardised by this. It emerges in the awareness of what the agency itself already knows.
For the senior layer in both house types, a second and equally significant effect appears. Experience can be deposited in a controlled environment in a form that reaches beyond the individual. A senior partner with twenty years of sector experience can keep notes, lessons learnt, methodological adjustments and pitch reflections in a Private AI Knowledge System in a way that remains retrievable for the generations that follow, without gathering dust in a classical knowledge management tool. The succession question changes its character. It shifts from a matter of personal mentoring to a matter of organisational design.
Across all three applications, the responsibility of the advisor remains intact. AI-supported preparation replaces neither strategic judgement nor creative invention. It makes both faster and better informed. That distinction is easier to maintain cleanly inside a controlled environment than inside an unmoderated reliance on public tools.
What a Private AI Knowledge System is explicitly not
The distinctions are worth drawing here too, because current market rhetoric tends to push every AI application in consultancies and agencies under the same umbrella. A Private AI Knowledge System is, firstly, not a pitch generator. It does not produce finished proposals at the press of a button. It makes the firm's pitch knowledge accessible so that the pitch team enters the room better prepared.
It is, secondly, not a substitute for style. A creative tonality, a brand voice, an argumentative manner emerges from a human engagement with material. A knowledge system makes the material broader and faster to reach. It does not replace the engagement, and should explicitly not attempt to.
It is, thirdly, not a control instrument. Anyone introducing a Private AI Knowledge System in order to monitor individual employees will watch it fail. Acceptance of such systems depends on visible relief for the staff and on a clear understanding that the firm is the responsible carrier of the infrastructure.
It is, fourthly, not a project to be handed to the IT department and signed off six weeks later. The question of what material, in what form, should be accessible to which roles is a strategic decision of the partnership or the management team. It concerns how the house treats its own intellectual substance.
What the market picture suggests
Reliable figures on the knowledge infrastructure of consultancies and agencies in the DACH region remain sparse, but the available picture is consistent. The Bundesverband Deutscher Unternehmensberatungen, the Gesamtverband Kommunikationsagenturen and comparable voices in Austria and Switzerland have for years described the same finding. The industry invests in methods, in talent and in tools. It invests comparatively little in the systematic condensation of its own knowledge. International work, including the annual Source Global Research report on the European consulting market and the SCOPEN studies on the agency landscape, points in the same direction. Houses that have invested early in their own controlled knowledge environment report, in their own perception, faster mobilisation of teams, shorter pitch preparation cycles and better handover quality during personnel changes.
These observations are not proof. They are an indication, and they point in a direction that has only recently begun to surface at the top of the houses concerned. The firms that take it early and seriously will, in five years, hold a body of condensed house knowledge that their slower-moving competitors will then have to reconstruct at considerable expense. Competitors without that body will lean instead on the body provided by public models, which is the same body equally available to everyone else in the room.
The question that remains
In most consultancies and agencies the AI conversation is currently framed as a productivity question. Which tools accelerate which steps, which licences ought to be procured centrally, which internal training will bring the staff up to speed. These questions are not wrong. They are merely secondary. The first question is what a consultancy or an agency actually intends to do with the thinking it has produced over the years. Whether it wishes to leave that thinking to the personal memory of its practitioners, whether it wants to bring it into a form that belongs to the firm as a whole, or whether it intends to entrust it to the same public models its staff is, in any case, already using without formal permission.
This question is strategic, it is commercial, and at heart it is a question of ownership. Anyone running a house whose principal asset is collective intellectual labour ought to confront it before the market does so on their behalf.
In the coming essays in this series, the same line of thought is extended to further segments. To engineering offices and technical service providers, to software companies, to industrial groups with distributed sites, and to the cross-cutting questions of governance and technical architecture. Consultancies and agencies appear early in this sequence because their knowledge is wholly immaterial. What they sell is thinking. What they lose, when that thinking does not remain in the house, is, quite literally, their business model.
