By the time a project reaches a formal dispute about scope, the conversation has usually moved on to contract clauses, variation orders and fee adjustments. The documents in front of both parties are the ones produced after the disagreement started.
The document that would settle it does not exist. It would be an accurate record of what the client said they wanted during the first two hours of conversation, before anything had been drawn, priced or approved.
What exists instead is a set of notes taken by the person who was also leading that conversation, plus two sketches and a memory that has since been reshaped by everything that happened afterwards.
What should a client brief include?
A project brief sets out what the client wants the project to achieve, within what budget, on what programme, and to what standard. The RIBA Plan of Work places this work in Stage 1, Preparation and Briefing, whose core tasks include preparing the Project Brief with project outcomes, sustainability outcomes, quality aspirations and spatial requirements, alongside feasibility studies and an agreed project budget. The stage outcome is a brief the client has approved and that has been confirmed as deliverable on the site. The full stage definitions are published in the RIBA’s own Plan of Work overview.
Read closely, that list contains an assumption worth surfacing. Every item on it originates in something a client said out loud, usually early, usually once.
Brief drift, and where it comes from
Scope creep is the familiar term and it points at the wrong culprit. It implies a client who keeps adding.
The more common failure is quieter. The brief as written diverges gradually from the brief as spoken, and neither party notices, because each individual step is small and reasonable.
The mechanism is consistent across project types. A client describes a requirement in their own words, which are imprecise but complete. The designer records a version of it, which is precise but partial. Work proceeds on the recorded version. Months later the client reviews a proposal and reports that it is not what they asked for, and they are frequently right, in the sense that what they originally said contained a qualification that never survived transcription.
The qualification is almost always one of three kinds.
The three things said early that rarely reach the brief
- Constraints stated as asides. A client mentions in passing that a relative uses a wheelchair occasionally, or that the property must be lettable for part of the year, or that a particular tree cannot be touched because of a family association. These arrive as conversational context rather than as requirements, and they are recorded as context or not at all. Each one is capable of invalidating a scheme.
- Priorities and the trade-offs behind them. Clients rarely rank requirements explicitly, but they signal ranking constantly through emphasis, repetition and the order in which they raise things. A written brief flattens all of this into a list of equal bullet points. When the budget later forces a choice, the ranking that would have made the choice obvious has been discarded, and the decision gets made by whoever is in the room.
- Prohibitions. What the client does not want is stated once, often negatively and often about a previous property or a previous architect. Negative requirements are the least likely to be written down and among the most likely to cause a rejected scheme, because they are invisible until violated.
None of these three are difficult to capture. They are difficult to capture while simultaneously running the meeting.
Why the person leading the meeting is the worst possible note taker
The first client meeting requires the designer to do several things at once: ask questions, watch reactions, form an early view of feasibility, manage the client’s expectations about cost, and build the relationship on which the commission depends.
Note taking competes with all of them. Attention spent on writing is attention removed from listening, and the specific casualty is the aside, which by definition arrives while the designer is dealing with something else.
There is a second problem, less discussed. Notes are selective by nature, and selection happens against the note taker’s current mental model of the project. A designer who has already formed a hypothesis in the first twenty minutes will, entirely in good faith, record the material that fits it. The client’s stray remark about lettability does not fit, so it is not written down. The hypothesis then goes unchallenged for months.
Bringing an assistant to take notes solves the attention problem and not the selection problem. The assistant selects too, against a weaker model.
Capturing the conversation as spoken
The alternative is to record the first meeting and produce the brief from a transcript rather than from notes.
This is not an argument for issuing a transcript to the client. A transcript is not a brief and should never be presented as one. It is a source document, in the same category as a measured survey: nobody hands the client a survey and calls it a design.
What changes is which stage the interpretation happens at. With notes, interpretation occurs in real time, under load, and is irreversible. With a recording, interpretation occurs afterwards, deliberately, and can be revisited when the project changes shape.
Three practical points govern whether this works.
Announce the recording at the start and record the fact that it was announced. A client who discovers afterwards that a commercial conversation was recorded without their knowledge has been given a reason to distrust everything that follows. State the purpose, which is accurate briefing, and state when the audio will be deleted. If the client objects, do not record.
Place the device where it can hear the room rather than in front of the designer. First meetings frequently involve two clients who interrupt each other, and the useful material is disproportionately in the interruptions.
Ask the closing question that most briefs lack. Something to the effect of: what would make you consider this project a failure. Answers to that question are the prohibitions, stated deliberately for once, and they are worth the recording on their own.
From recording to brief
The extraction is a short editorial task, not a transcription task.
Converting the audio is the mechanical part. A tool that will turn speech into text from a recording made in the browser or uploaded from a phone will return punctuated text with the speakers labelled, which matters in a meeting with two clients whose requirements do not always agree. Vomo also produces timestamped chapters and a summary with action items, so a two hour conversation arrives already divided into sections that can be worked through in order.
Four passes over that text produce a defensible brief.
First, extract every stated requirement verbatim, with its timestamp. Verbatim matters here; paraphrase at this stage reintroduces the original problem.
Second, extract every constraint, including the asides. These are identifiable by their form: they are usually subordinate clauses about people, seasons, money or history.
Third, extract every prohibition, and list them as prohibitions rather than folding them into positive requirements.
Fourth, and only then, write the brief in professional language, with each item traceable to the point in the conversation it came from.
The traceability is the deliverable. A brief in which every line can be sourced to a timestamped statement by a named person is a different instrument from a brief that represents the designer’s best recollection.
The confirmation step
The brief should then be returned to the client with an explicit request to correct it, and the correction round should be treated as a stage in its own right rather than a formality.
Clients read a written version of their own requirements differently from how they said them. Reading their intentions back to them in professional language routinely produces new information, because the written form makes gaps visible that speech conceals. This is the cheapest moment in the entire project to discover a requirement.
Once the brief is signed off, the audio has served its purpose and should be deleted on the schedule stated at the outset. Private client work frequently involves confidential information about family circumstances, finances and security arrangements; retaining recordings beyond the point of use creates an obligation with no corresponding benefit. Where a practice handles such material, the storage location, retention period and access rights should be settled once, in writing, and applied to every commission rather than decided project by project.
The narrow claim
None of this prevents clients from changing their minds, which they are entitled to do and frequently should.
What it prevents is the specific and expensive situation in which both parties are certain about what was agreed at the outset, both are arguing in good faith, and no record exists that is capable of resolving the question. That situation is not caused by difficult clients. It is caused by asking one person to lead a two hour conversation and document it at the same time, and then treating the output as authoritative.
