Client Brief Generator for Agencies: The Smart Intake Form That Chases Every Missing Answer
Every agency has a client folder that looks like this. A signed contract. A deposit paid. And an intake form where the target audience says "everyone," the deadline says "ASAP," and the logo files are "coming Monday." That was 3 Mondays ago.
The project can't start. The account manager has sent 4 polite emails, each one a little less polite. The designer is booked for a job that has no brief. And when the brief finally does get written, half of it is guesswork the client will push back on in round 2.
This post walks through a tool any agency can build in a couple of days to fix that: a client brief generator that sits behind a smart intake form, checks every answer with AI, and keeps chasing the gaps on its own until the client either finishes or tells it to stop.
Want the full build? The Smart Intake Form & AI Client Brief Generator with Gap Chaser pro prompt has the complete Lovable master prompt (24,000 characters, ready to paste), a 10-step build guide, an 8-step sell-and-run playbook, pricing and a cold outreach script.
Where agency onboarding actually breaks
Ask any agency owner where projects slip and you'll hear the same 3 stories.
The first is blanks. The client skipped the questions they didn't feel like answering, usually the hard ones. Budget. Competitors. Who signs off.
The second is vague answers, and these are worse than blanks because they look finished. "Our audience is everyone." "We want it modern and clean." "More sales." A Google Form happily accepts all of them. Then a strategist spends the kickoff call trying to drag a real answer out of the client, or worse, writes the brief around a guess.
The third is chasing. Somebody has to remember which client still owes what, write a fresh email each time, and keep doing it without sounding annoyed. In practice that person is an account manager with 12 other clients, so the chasing happens in bursts and then stops.
All 3 are plumbing problems, and plumbing is exactly what a well-built tool fixes.
What the tool does, end to end
Here's the flow from the agency's side.
The account manager creates a new intake, picks the services the client bought (say, website design and paid social ads), and hits send. The client gets an email with a private link, plus a text if they ticked the SMS consent box. No account to create, no password to forget.
The link opens a branded form in the agency's colours. It only shows the questions that matter for the services on this job. A website client gets asked about their CMS, page list and who's writing the copy. A paid ads client gets asked about monthly ad spend, the offer and whether the tracking pixel is installed. Everyone gets the shared basics: goals, audience, brand, budget, deadline, approvals.
Files go straight in. Logos, brand guidelines, photos, PDFs. Anything over 50 MB gets a "paste a link instead" box for Google Drive, Dropbox or WeTransfer. Everything autosaves, so a client can start on their phone in a taxi and finish at their desk.
The quality check
This is where it earns its keep. After each answer, AI checks it against two things: is it one of the agency's mandatory fields, and does it meet the quality bar for that question?
Each question in the bank carries a plain-English quality rule. For target audience, the rule says the answer has to name a specific group by role, age band, location or situation. So when the client types "everyone," they see a friendly amber note under the field within a couple of seconds:
"We'll need something more specific here. Who buys most often: homeowners, landlords or businesses? Roughly what age and where?"
The client isn't blocked. They can move on. But that mandatory field stays flagged as weak, and the intake stays incomplete until it's fixed or the account manager accepts it after a phone call.
The same logic catches "ASAP" as a deadline, "more sales" as a goal with no number attached, "modern, clean, professional" as brand words, and "we have no competitors" (every client has competitors; they just haven't looked).
The gap chaser
Now the part most intake tools skip. Anything still missing or weak gets chased automatically.
The first chase goes out 48 hours after the form was sent (or 48 hours after the client last touched it). If that chase gets no response, the wait doubles: 4 days, then 7 days. Then it holds at 7 days and keeps going. Every week. Indefinitely.
That's deliberate. Most drip sequences send 3 reminders and give up, which means the 4th week of silence is exactly when the client drops off everyone's radar. This one only stops for 4 reasons: the intake is complete, the client opts out, staff pause it (say the client is on holiday until the 14th), or the intake gets archived.
And each chase names the exact items still outstanding, something like:
"You're 80% done. 3 quick answers and we can start: who your main customer is, your launch date, and 2 competitor websites."
Each item links straight to that question in the form, with the AI's follow-up question attached. It alternates between email and SMS where the client has opted in to texts, never sends during quiet hours or weekends, and rotates its wording so the 6th chase doesn't read like the 1st.
If a client ignores 3 chases in a row, the account manager gets a red "needs a call" flag on the dashboard and an email with the client's phone number. The automatic chasing carries on in the background. The human gets pulled in on top of it.
The brief and the sign-off
The moment every mandatory field passes, the app writes the creative brief in the agency's own template: project overview, objectives and success measures, audience, proposition, brand and tone, scope, assets, budget and timeline, approvals. Each section shows small chips linking back to the client answers that fed it.
Where the template needs something the client never said, the AI writes it as an explicit assumption in a separate list. It doesn't invent facts. It also suggests up to 5 open questions for the kickoff call.
The account manager edits, approves internally and sends it to the client. The client reads it as a clean document, then either requests changes (per section or overall) or approves it by typing their full name and ticking a confirmation box. The app records the name, time, IP address and browser. That record matters later, when someone in month 3 says "we never agreed to that."
The approval request gets chased on the same doubling cadence as the form, so a brief doesn't sit unread in someone's inbox for 2 weeks either.
If you want to see how the master prompt specifies all of this (tables, screens, the chase rules, webhooks and the edge cases), it's laid out in full in the client brief generator blueprint.
A worked example: Jo and the bakery rebrand
Let's run a realistic client through it.
Northlight Studio, a 9-person agency, signs Jo, who owns 2 bakeries and wants a brand refresh plus a new website. Sam, the account manager, creates the intake on Tuesday morning with brand identity and website design selected, and adds Jo's business partner Priya as a second contact responsible for the Brand section.
Jo opens the link on her phone that evening. She fills in the business basics, uploads 3 old logo files, and types "everyone who loves good bread" for the target audience. The amber note pops up asking whether her regulars are mostly local families, office workers or weekend visitors. She ignores it for now. She leaves the deadline blank and closes the tab. The dashboard shows 45%.
Thursday evening, 48 hours later, Jo gets a short email listing 4 items: target customer, launch date, current website login, and who signs off. Each one is a link. She fixes the audience ("local families within 3 miles, plus office lunch trade on weekdays"), sets the launch for 1 March because of a new shop opening, and stops again. 75%.
Priya, meanwhile, hasn't opened her link at all. Her chase arrives by text on Friday, since she ticked SMS consent. She answers the brand questions from her phone on Saturday morning, picking "warm, floury, a bit cheeky" as the brand words. The AI passes that immediately. "Cheeky" is distinctive.
The website login is still missing. The next chase to Jo goes out the following Wednesday, 4 days on. She replies to the email asking whether they need the hosting login or the WordPress one. The reply lands with Sam, because every email's reply-to is the agency's own inbox. Sam answers, Jo adds both, and the intake hits 100% on Thursday.
Sam gets a push notification on his phone (more on that below), opens the generated brief, tightens 2 sections, and sends it. Jo requests one change: she wants the second shop's opening on the timeline. Version 2 goes out the next morning. Jo approves it with her typed name at 7:42 on a Saturday.
Total from contract to signed brief: 11 days. Manual follow-up emails Sam had to write: 1, and that was answering a question. Under Northlight's old Google Form process, Sam's own estimate was 3 to 4 weeks with somewhere around 8 to 10 back-and-forth emails.
How it gets built
The blueprint is built for Lovable, with Lovable Cloud handling the database, file storage, scheduled jobs and server functions, and Lovable AI doing the quality checks and brief writing. Resend sends the emails from the agency's own domain. Twilio sends the texts and listens for STOP replies.
The master prompt is one paste. It tells Lovable not to ask clarifying questions and lays out the whole application: about 15 tables, 4 main screens, the quality check and brief generation functions, the 15-minute chase job, the Resend and Twilio webhooks, a reports page, and a seeded question bank so the app works on day one. It also seeds 6 demo clients at different stages, which makes the sales demo look like a real agency's dashboard instead of an empty screen.
You fill in 8 fields before copying it: agency name, the services they sell, brand colour, lead account manager, reply-to email, time zone, hours before the first chase and the longest gap between chases. Everything else is customised afterwards in the app's own settings, with no code. Question wording, mandatory flags, quality rules, branching, brief templates and chase message wording all live in admin screens.
The build guide on the pro prompt page covers each step click by click, including verifying the sending subdomain in Resend, buying a Twilio number, and testing the chaser without waiting 2 days (set the first-chase delay to 0, wait for the next 15-minute run, and check your own phone).
The SMS paperwork
One thing to plan for. US carriers block unregistered business texting, so before SMS chasers go live the agency needs A2P 10DLC registration through Twilio. For most small agencies that's a $4.50 brand registration (or $46 for a Standard brand with secondary vetting), a $15 campaign vetting fee, and a monthly campaign fee between $1.50 and $10. Approval can take 1 to 3 weeks.
The app runs on email-only chasing until then, so this doesn't hold up launch. It just means texts switch on a little later.
Shipping it as a real app
The last step wraps the dashboard as a native iOS and Android app (plus a PWA) through AppBuild.DIY. Push notifications go on so account managers hear the moment a client finishes, approves, asks for changes or gets flagged for a call. An email alert gets buried under 40 others; a push on the lock screen gets read in minutes. The Photos/Camera plugin goes on too, so staff can snap whiteboard notes or a client's printed samples straight into an intake.
There's a sales angle as well. An agency that onboards clients through its own branded app looks bigger than it is, and that's a story it can tell its own prospects.
Who should build this, and the money
There are 2 kinds of buyer here.
The agency building it for itself. The running costs are small. Lovable Pro is $25 a month. Resend is free up to 3,000 emails a month, and $20 a month for 50,000 if you outgrow that. Twilio SMS starts at $0.0083 per US message plus carrier fees, so even a busy month of chasers is a few dollars. Compare that with off-the-shelf options. ScopePilot, which turns intake answers into briefs, is €19 a month for 15 briefs or €49 for unlimited. Content Snare, which handles content collection with reminders, runs $35 to $119 a month billed annually. Both are decent tools. Neither asks your exact questions, writes the brief in your exact template, or chases indefinitely on your rules. ScopePilot, for example, fills gaps with editable defaults, and this tool chases the client for the real answer.
The freelancer or studio selling it to agencies. This is where the pro prompt's pricing comes in:
- Setup: $2,500 to $4,500, covering the build, customising questions and templates for up to 5 services, connecting live accounts, the A2P paperwork and a supervised 3-client pilot.
- Monthly retainer: $197 to $397, covering hosting, AI and messaging costs at normal volumes, monthly reporting and up to 2 hours of tweaks.
- Performance option: $147 a month plus $25 per brief approved through the tool. At 12 approved briefs a month, that's $447.
- Multi-brand: an extra $99 a month per additional brand sharing the build.
Why would an agency pay that? Because its own projects are big. Small agencies typically quote $3,000 to $15,000 for a custom website, and branding or ad retainers sit in similar territory. If the tool pulls a single project's start date forward by 2 weeks, or saves one round of revisions caused by a brief built on guesses, it has covered a year of the retainer. Most agency owners can name a project from the last quarter that slipped because a client went quiet.
Finding buyers isn't hard either. Agency directories like Clutch and DesignRush list thousands of small studios by city. Agency-owner groups on Facebook and LinkedIn are full of people complaining about exactly this. And any agency whose "start a project" page leads to a generic questionnaire is already showing you the problem.
The sales script in the blueprint opens with one question: how many days did your team spend last month waiting on clients? It's a good opener because every agency owner knows the answer and none of them like it.
What to expect in the first month
Be honest with buyers about this.
The build takes 1 to 2 days for someone comfortable with Lovable. The first 2 weeks of the pilot involve tuning, because the AI will be a little too strict on some questions and a little too relaxed on others until the quality rules get adjusted. That's normal, and it's why the playbook includes a supervised pilot with the agency's next 3 clients.
The strategy conversation still happens. It starts from a complete, specific set of facts, so the kickoff call is about ideas instead of chasing a logo file.
And some clients will still go quiet. The difference is that the system notices within days, keeps nudging, and tells a human when it's time to pick up the phone.
The Reports page closes the loop. It shows average days to complete, chases per completed intake, first-version approval rate, and the 10 questions clients most often answer vaguely. That last list is gold. If "describe your brand personality" gets flagged on 70% of intakes, the problem is the question, and now the agency knows to rewrite it.
Part of a bigger set
This is the general-purpose intake engine in Prompt-King's new Agency Systems collection. If your agency specialises, the sibling blueprints go deeper on one brief type: the SEO content brief generator, the UX design brief generator and the brand identity brief generator. They pair well with this one. The intake form collects the facts, and a specialist brief builder turns them into the deliverable-specific document.
Build it this week
Every agency already has an intake process. It's just held together by an account manager's memory and a folder of half-finished Google Forms. Swapping that for a form that flags weak answers on the spot, chases every gap until it's closed, and hands back a signed-off brief is one of the fastest wins an agency can get, and it pays for itself on the first project that starts on time.
Get the full blueprint: the Smart Intake Form & AI Client Brief Generator with Gap Chaser includes the paste-ready Lovable master prompt, every build and sell step, tool costs, pricing tiers and a cold email you can send today.