The agency intake form that collects everything and tells you nothing

I have sat in kickoff calls where the account lead asks the client what their target audience is. The client filled in a Typeform three days earlier with four paragraphs on exactly that. Nobody opened it before the call.

The intake form was not the problem. The absence of any step between "form submitted" and "meeting scheduled" was.

This happens in agencies that have genuinely invested in their intake process. The Typeform is conditional-logic'd to within an inch of its life. The Notion template has toggle sections and linked databases. Someone spent two afternoons on it. And it sits, unread, in a submissions inbox while the account manager books the kickoff off memory and instinct.

What the intake form problem actually is

The standard diagnosis is that the form is incomplete. So agencies add fields. Budget range. Timeline. Preferred communication style. Do you have brand guidelines? Please upload them. Have you worked with an agency before? If yes, what went wrong?

The form gets longer. The response quality stays roughly the same. Scope creep continues. Post-project retrospectives still surface "we didn't fully understand what they wanted at the start."

The form was never the bottleneck. The bottleneck is the gate that should exist between submission and the first human conversation. That gate is missing. Or rather, it exists in theory and nobody is standing at it.

What agencies actually need is not a better intake form. They need a triage step. A defined moment where someone, or something, processes the submission before the kickoff is booked. That moment does not exist in most agency workflows. The form submits, a notification fires to Slack or email, and the next action is a calendar invite.

The funnel has a mouth and it has an output. The middle, where the intake data gets converted into a qualified brief, is a gap.

Why completeness made things worse

There is a specific failure mode that comes from thorough intake forms. When a form has thirty fields, the person reading it faces a choice: read all of it, or skim it and trust that they will catch anything important in the meeting. They skim it. The meeting then functions as a second intake, which means the form was a formality.

Worse, when the form is long, clients write less carefully. Short forms with focused questions tend to produce considered answers. Long forms produce fatigue. By field eighteen, the client is typing "TBD" and "see previous answer."

The form length also creates a false sense of preparedness. The account lead sees a full submission in the inbox and assumes the brief is solid. They go into the kickoff feeling informed. They are not. They have unread information that may contradict what the client says in the meeting, and no process for surfacing the contradiction.

I wrote about a version of this in the context of AI agency briefs in the brief bottleneck piece. The same dynamic applies here. More input fields do not produce better briefs. A defined processing step does.

The triage step that most agencies skip

A triage step is a fixed action that happens after form submission and before kickoff. It takes the raw intake data and converts it into something a human can act on in a meeting. It does not have to involve AI. It does have to involve someone.

In a small agency, this might be fifteen minutes with the form responses, a short internal summary document, and a list of three to five questions to resolve in the kickoff. The summary does not repeat the form. It identifies what is clear, what is ambiguous, and what is missing. The kickoff then has an agenda built from that list.

In a larger agency, or one processing volume, this is where AI summarisation earns its place. A language model can read a form submission and produce a structured brief summary in under a minute. It can flag contradictions (the client says "we want to grow brand awareness" and also "we need leads by end of Q3"), missing fields, and responses that are vague enough to need clarification. That output goes to the account lead before the kickoff is booked.

The AI does not replace the meeting. It replaces the skim.

What matters is that the triage step is assigned. Someone owns it. It appears in the project management tool as a task between "intake received" and "kickoff scheduled." Without that, even the best AI summarisation output will sit unread in the same way the form does now.

What the workflow actually looks like

Here is a minimal triage workflow that works with or without AI tooling:

intake_triage_workflow:
  trigger: "Client intake form submitted"
 
  step_1:
    name: "Auto-acknowledge to client"
    owner: "system"
    action: "Send confirmation email with expected next-steps timeline"
    timing: "Immediate"
 
  step_2:
    name: "Triage task created"
    owner: "system"
    action: "Create task in project management tool assigned to account lead"
    timing: "Immediate"
    task_title: "Review intake and prepare brief summary before booking kickoff"
 
  step_3:
    name: "Brief summary produced"
    owner: "account lead"
    action_without_ai: "Read full submission, write 1-page summary covering: what is clear, what is ambiguous, what is missing, 3-5 questions for kickoff"
    action_with_ai: "Paste submission into brief-summary prompt, review output, edit and add context, confirm questions list"
    timing: "Within 24 hours of submission"
    output: "Internal brief summary doc linked to project record"
 
  step_4:
    name: "Kickoff gating check"
    owner: "account lead or ops"
    action: "Kickoff is not scheduled until brief summary exists and has been reviewed"
    timing: "Before calendar invite is sent"
 
  step_5:
    name: "Kickoff with agenda"
    owner: "account lead"
    action: "Meeting agenda built from questions in brief summary, not from memory"
    timing: "After triage complete"

The gate in step 4 is the part most agencies do not have. The kickoff can be booked by anyone with calendar access. There is no check. Adding that check, even as a manual confirmation, changes the behaviour of the whole workflow.

Where to start

If you are running an agency and recognising this pattern, the fix is not a new intake tool.

First, audit one month of kickoffs. For each one, check whether a brief summary document existed before the meeting was booked. Not the form submission. A processed summary. My guess is the number is low.

Second, write the brief summary prompt if you are using AI, or the summary template if you are not. One page. Four sections: what we know, what is unclear, what is missing, questions for kickoff. Test it on three recent submissions.

Third, add the triage task to your intake automation. If your form tool is Typeform, Jotform, or a similar platform, it almost certainly has a Zapier or Make integration. The trigger is form submission. The action is a task created in ClickUp, Asana, Linear, or wherever your team works. Assign it. Give it a due date.

Fourth, add the gate. The kickoff invite does not go out until the triage task is marked complete. This can be a manual check or an automated dependency. Either works. What matters is that it exists.

The form you have is probably fine. The workflow around it is where the brief is actually getting lost. That is also where scope creep starts. Not in the project itself, but in the gap between what the client submitted and what the team understood before the work began.

If you want a structured look at where your intake workflow is losing information before it reaches the people doing the work, the AI Workflow Audit covers exactly this: where the gates are, whether anyone is standing at them, and what a triage layer would look like in your specific setup.