Thousands of reviews nobody read became a brand's monthly source of new ads
Adam Taylor's team loaded Claude with every review a supplement brand had ever received, pulled customers' exact words and tagged every ad by angle, persona and hook. He says that loop is why his strategists find new winning angles every month.
Published Oct 26, 2026
The Wow
What someone built
A creative-agency method: verbatim review mining with Claude, every angle and script starting from customer language, and an ad log tagged by angle, persona and hook so each launch shapes the next.
- Original source
- Adam Taylor's review-mining workflow (X, Oct 25, 2026)
- License
- A described method; we quote and link the post
- To use it
- Nothing to install: it's a method. The prompt below sets it up in any AI chat.
- A business you could build on it
- Monthly review-to-angle desk
Why it's remarkable
The tools that mine reviews already exist. The compounding comes from tying customer words to the brand's own ad results, month after month, and that loop is a service.
The Opportunity
The business it reveals
A monthly review-to-angle brief for DTC brands and small creative agencies, priced between one-off review analyses and an in-house strategist.
Who would pay
DTC brands spending on Meta or TikTok ads, and agencies whose strategists need fresh angles.
The Flip
3 ways you could turn this into a business
Monthly review-to-angle desk
- Who pays
- DTC brands running Meta or TikTok ads
- How it would earn
- A monthly fee for an angle brief from new reviews plus an updated ad log (our hypothesis; one-off review analyses list at about $300–$400)
- Why it might work
- Brands already buy one-off review analyses and self-serve tools; a recurring brief tied to their own ad results is the step neither offers.
- What already exists
- Self-serve tools such as Adlicio and AdMuse ($29–$199 a month list), one-off analyses such as WhyHire (from $299) and in-house strategists.
Review-to-landing-page copy for SaaS
- Who pays
- B2B SaaS teams with G2 or Capterra reviews
- How it would earn
- A per-page fee to rewrite headlines and objection sections in customers' own words (our hypothesis)
- Why it might work
- SaaS reviews are long and specific, and landing-page tests are cheap to run.
- What already exists
- Conversion copywriters who do review mining by hand and AI copy tools.
Google-review insight report for local chains
- Who pays
- Restaurant groups, clinics and service chains with several locations
- How it would earn
- A quarterly report on what customers praise and complain about by location (our hypothesis)
- Why it might work
- Owners read a few reviews but never the pattern across hundreds, and each location manager wants their own list.
- What already exists
- Reputation-management software and mystery shoppers.
Why we'd try "Monthly review-to-angle desk" first: The buyers already spend on ads every month, the inputs are exports they own, and the tagging loop makes your work more valuable each month instead of easier to replace.
Build It
How you could build it
A monthly brief per client: top customer phrases by pain, outcome and objection, five new angles with the exact quotes behind them, and an ad log tagged by angle, persona and hook that shows what won last month.
First step: Pick a DTC brand with a few hundred public reviews. Paste 200 of them into an AI chat with the prompt below and see whether the five angles it returns are ones the brand isn't already running in Meta's Ad Library.
Agents and tools
- An AI agent with a large context window (ChatGPT, Claude, Gemini, Grok Bot or similar)
- The client's review export
- The client's ad results export
- A spreadsheet ad log
Setup
- Collect the client's review export and the last 90 days of ad results.
- Run the mining prompt and check every quote against the source.
- Tag past ads by angle, persona and hook in the sheet.
- Write five angles, each tied to quotes and to past results.
- Repeat monthly with new reviews and new results.
What the agent does
- Reads every review
- Pulls exact customer phrases
- Clusters pains, outcomes and objections
- Drafts angle briefs
What you do
- Gets exports from clients
- Checks quotes and claims
- Chooses angles with the strategist
- Keeps the ad log honest
Your first customer: A brand with lots of reviews and tired-looking ads, offered a free first brief from its public reviews.
Test demand first: Send a free brief to five brands. Success: at least two reply, and at least one runs an angle from it or asks for a monthly price.
Your starter instructions
Paste this into the AI agent you use (ChatGPT, Claude, Gemini, Grok Bot or another) as your first message. Copying it starts work in your own agent; it doesn't create or deploy a business by itself.
Running it
What stays human: Getting review and ad exports from clients, checking every quote against the source, choosing which angles to test, and writing or approving the ads.
- Startup cost (estimate)
- About $20–$100 to start (estimate) (Estimate, not measured: an AI plan with a large context window and a spreadsheet; clients supply their own exports)
- Running cost
- An AI plan, tens of dollars a month (estimate) (Rough estimate: monthly re-runs on each client's new reviews; not measured)
- How much the agent runs
- People run it; AI helps with tasks.
Pricing your own version? Use the free startup-cost worksheet.
The evidence
What the source shows: Adam Taylor's X post (Oct 25, 2026), describing his agency's method. No results data is shared. Tool and service prices are their published list prices.
What's ours: The desk, its pricing and the flips are our application.
Worth planning around
- One agency's description of its own method.
- Supplement and health claims in ads are regulated; customer words still need checking before they run.
Sources
- Adam Taylor: review-mining workflow (X, Oct 25, 2026)
- Adlicio pricing (review-mining tool)
- WhyHire Shopify review and voice-of-customer analysis (pricing)
Sources reopened Oct 26, 2026. Spot a mistake? Tell us and we'll correct it in public.
Keep reading
- How much of the work do AI agents actually run? (guide)
- He still can't write code. His five-person ad agency now runs on about 270 tools he built with Claude. Every owner-run firm needs its first five
- A marketer who doesn't code built an elk-call coaching app with Claude Code and says it hit a $650K run rate in 110 days. Every skill learned by sound is the same opening
- A free n8n workflow flags fake supplier invoices before anyone pays them — and Germany's e-invoice switch makes it a service bookkeepers can sell
- Marc Lou's agent chased a broken-shoe refund through three support desks and got his 850 AED back. Small stores now need a claims desk that can answer agents
- Every discovery and launch plan