Creative Brief Builder

Two ways in, one conversation. Give it a product URL and it scrapes the page and builds the full brief pack. No URL? Paste what you know about the product and audience, and it produces the same outputs without the scrape step.

Every run starts with the emotional truth: the specific feeling your product resolves, written in your customer's own words rather than marketing language. The agent gets there by asking five plain questions about the customer experience. Everything else in the run traces back to that answer.

The image pack is 20 briefs, organized so you test the four angles that matter: the problem told through feeling, the problem told through logic, the outcome told through feeling, the outcome told through logic. Each brief comes with scored headlines, a ready-to-paste AI image prompt for Midjourney, DALL-E, or Flux, and exact specs for Meta, TikTok, and YouTube.

The video pack is three to five briefs across formats that are proven to work - mechanism explainer, founder story, comparison, testimonial compilation - each with its hook scored against eight direct-response principles.

And if you're running YouTube, it plans the campaign around the creative too: budget splits across prospecting, retargeting, search, and testing, plus a 10-week calendar for scaling what wins.

Emotional truth is the specific customer feeling the product resolves, written in customer language rather than marketing copy. The agent extracts it from a five question discovery, framed in plain conversational terms: what was the moment they realized they needed this, what failed before, what does the resolution actually feel like, what would they say to their past self, what would they say to a friend.

Every downstream output, including hooks, scripts, image briefs, and thumbnails, traces back to the same emotional truth. That is why creative built through this agent reads coherent across channels instead of looking like four different freelancers wrote it.

Twenty briefs structured across four angle archetypes, with five briefs per angle: Problem and Emotional (the pain framed by feeling), Problem and Rational (the pain framed by logic), Outcome and Emotional (the after, framed by feeling), Outcome and Rational (the after, framed by logic).

Each brief includes scored headlines, a ready-to-paste AI image prompt for Midjourney, DALL-E, or Flux, and clear visual rules: one dominant element that owns the canvas, no more than four text blocks, copy and visual telling the same story, a seven-word simplicity test, and a does-it-read-as-a-thumbnail check. Platform specs for Meta, TikTok, and YouTube come attached to every brief.

A four layer budget and audience structure. Layer one is prospecting, where 60 to 70 percent of budget goes against cold audiences. Layer two is retargeting, where 10 to 15 percent goes against warm audiences.

Layer three is search capture, where 20 to 25 percent goes against intent. Layer four is a testing lab, where 5 to 10 percent runs isolated tests.

The staging sequence is concrete. Launch layer one and layer four together. Add layer three (search capture) after three weeks. Add layer two (retargeting) after three months, once remarketing pools have data.

Bidding transitions follow a fixed order: Max Conversions first, then target cost per acquisition once you have 30 conversions, then target return on ad spend after another 30.

Creative graduation thresholds: above 1.8 return on ad spend graduates to layer one, 1.5 to 1.8 iterates once, below 1.5 is killed.

Fatigue triggers fire when click through rate drops more than 15 percent, frequency rises above 4.0, conversion rate declines more than 20 percent, or view rate drops more than 10 percent.

No. By design. The agent ships the brief and the prompts. You run the actual generation through Midjourney, Runway, Flux, your in house tools, or a designer or creator.

The reason is simple. The bottleneck for performance creative is not pixel generation. Every tool on the market can do that now.

The bottleneck is brief quality: which angle, which emotional truth, which hook, what proof, what visual hierarchy. That is where the agent compounds, and that is what it is paid to do.

The hosted website agent uses the data it can actually access: your product URL, pasted customer notes, uploaded files, brand assets found on the page, and optional Meta ad pattern research when available. It labels what is verified, inferred, assumed, or missing so you know how much confidence to place in each brief.

If you want it sharper, paste reviews, post-purchase survey comments, support tickets, creator notes, or competitor examples into the intake form or chat. The agent will use those as source material and mark them as user-supplied rather than inventing customer language.

Eight principles, each scored zero to ten: specificity, visceral language, tension and relief, loss framing, borrowed authority, proof integration, one clear action, and continuation.

Paste an existing Demand Gen ad or YouTube ad and the agent scores it across the eight, identifies the lowest scoring principles, and tells you exactly which sections to rewrite. Most ad fatigue is not audience fatigue. It is structural weakness in two or three of these principles.

A composite score above 60 ships. Below 40, rewrite from emotional truth. Between those two, iterate on the bottom two principles first.

Yes. There is a creator brief sub mode. Ask for a creator brief or a user generated content brief and the agent shifts. It scripts hooks only, then leaves everything else as talking points and guardrails, because over scripting kills the organic feel that makes creator content convert.

Output ships in two formats: a creator facing markdown brief (what you send the creator) and a machine readable JSON file (what feeds downstream production tools). Archetypes covered include yapper, demo, testimonial, founder, group, and discovery.

It does not build search or Performance Max campaigns or rewrite Shopping feeds. Hand off to Google Ads Agent for that work.

It does not build the landing or presell pages the creative points at. Hand off to Landing Page Builder. It does not generate the actual creative pixels. The output is brief, prompt, and specification.

It cannot replace a creative strategist's judgment in heavily regulated categories such as medical advice, FDA constrained supplements, or regulated financial products. Bring a human reviewer for compliance review in those categories.

A week by week deployment that integrates creative refresh cycles with campaign scaling milestones.

Week one launches with layer one (prospecting) and layer four (testing lab) on Max Conversions. Week three adds layer three (search capture). Week six transitions layer one to target cost per acquisition after the first 30 conversions clear.

Week eight graduates winning creatives, kills losers, and refreshes briefs from the same emotional truth. Week ten retests with the second creative wave.

Alongside that, the calendar specifies fatigue triggers, budget scaling decisions (when to double down versus hold), and the middle performer protocol (what to do with ads that are not winning but are not losing either).