Who does ChatGPT recommend? Sell the answer as a report
Search Console can't tell a business whether AI assistants recommend it. One SaaS founder measured it with 25 buyer questions on two engines, published the method and the scoring code, and found a competitor on the shortlist 19 times out of 50.
Published Oct 23, 2026
The Wow
What someone built
WEBVOGUE Technologies collected 50 live answers (week of Oct 12, 2026), logged every brand named, its position and every cited URL, and scored three numbers: visibility share, citation share and average position. The author builds both the product measured and the tool he used, and discloses it.
- License
- Article and code snippets shared publicly by the author; no license is stated, so follow the method and write your own code rather than copying it
- What you need
- Access to ChatGPT and Gemini with web search
- A list of 25 or so unbranded buyer questions
- A spreadsheet, or the author's scoring code
- Install
- Read only: the post walks through the method, the scoring function and the rules that keep the numbers honest. You can start by hand with a spreadsheet.
- A business you could build on it
- "Does AI recommend you?" reports for local service businesses
Why it's remarkable
His site was cited as a source more often than his product was recommended, and the single most cited source was a competitor's own website. The whole scoring logic fits on one screen.
The Opportunity
The business it reveals
Plain-English "does AI recommend you?" reports for local service businesses, with monthly re-runs. AI-visibility tools are built for marketers; the dentist or HVAC owner who'd most want the answer has nobody showing it to them.
Who would pay
Local service businesses in one niche and city, SEO agencies that want white-label reports, and SaaS marketing leads before a launch.
The Flip
3 ways you could turn this into a business
"Does AI recommend you?" reports for local service businesses
- Who pays
- Dentists, law firms, med spas and home-service companies in one city
- How it would earn
- A fixed-price first report plus monthly tracking (our hypothesis; price untested)
- Why it might work
- Owners can't see this anywhere else, the report is easy to understand, and a competitor's name in the answer creates urgency.
- What already exists
- AI-visibility tools such as Otterly.ai, Profound, Peec AI and the author's CiteGaps target marketers; local SEO agencies are starting to add it. The edge is a plain-English report for owners in one niche.
White-label AI share-of-answer reports for SEO agencies
- Who pays
- Small SEO and content agencies
- How it would earn
- A per-client monthly fee for branded reports (our hypothesis)
- Why it might work
- Agencies' clients are asking about AI search, and a ready monthly report lets an agency answer without buying another platform.
- What already exists
- The same AI-visibility platforms with agency plans; the edge is done-for-you analysis rather than another dashboard.
Competitor AI-answer audits for SaaS launches
- Who pays
- Seed and Series A SaaS marketing leads
- How it would earn
- A one-off audit with a content fix list (our hypothesis)
- Why it might work
- The post shows the fixes come straight from the data: get onto the pages engines cite and answer the questions you're missing.
- What already exists
- In-house marketers with the same tools, and SEO consultants adding AI answers to their audits.
Why we'd try ""Does AI recommend you?" reports for local service businesses" first: The tools are built for marketers; the local owner who most needs the answer has no one telling them, and one screenshot of a competitor in the answer sells the report.
Build It
How you could build it
A repeatable report: a fixed question set per niche, live answers with sources, the three scores, the pages the engines rely on, and five fixes.
First step: Write ten unbranded questions a customer would ask in one local category, ask each in a fresh ChatGPT and Gemini chat, and log who's named and which pages are cited.
Agents and tools
- ChatGPT and Gemini (by hand or via official APIs with web search)
- A spreadsheet or small script
- An AI agent to adapt prompts and check the work (ChatGPT, Claude, Gemini, Grok Bot or similar)
Setup
- Pick one niche and one city.
- Write 15 to 25 unbranded buyer questions grouped by intent.
- Collect answers in fresh sessions and log brands, positions and cited URLs.
- Compute visibility share, citation share and average position for the top ten businesses.
- Turn the gaps into a one-page report with five fixes.
What the agent does
- Extracts brands and citations from each answer
- Computes the scores and the trend
- Drafts the report and the fix list for your review
What you do
- Writes and fixes the question set
- Checks the extraction on a sample
- Presents the report and sells the plan
- Never promises placement
Your first customer: A local business that a competitor beats in your ten-question sample.
Test demand first: Run free ten-question samples for ten local businesses and send each its result. Success: two ask for the full report. Failure: no replies, which says owners don't feel the problem yet.
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: Writing the buyer questions with the client, reading the answers, turning gaps into a fix list, and selling and presenting the report.
- Startup cost (estimate)
- About $0 to $50 to start (estimate) (Estimate: the method works by hand with a spreadsheet; automating it adds API usage for each answer collected. Not measured)
- Running cost
- A few dollars a month per client (estimate) (Estimate: API usage for 50 or so answers a month plus extraction calls; check current API pricing. 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: The author's DEV post with method, results and scoring code (Oct 23, 2026). We read it; we haven't run the method ourselves.
What's ours: The local report service, pricing and flips are our application.
Worth planning around
- Fifty answers is a small sample; one answer moves a score by 2 points.
- The author says his question set mixed buyer intents, which widened the gap.
- Use official APIs or collect by hand rather than scraping chat apps.
- Dedicated tools already exist, so the edge is the niche and the advice.
Sources
- WEBVOGUE Technologies on DEV: I asked ChatGPT and Gemini 50 questions my buyers ask (Oct 23, 2026; author discloses he builds the product measured and the tool used)
- WEBVOGUE Technologies on DEV
Sources reopened Oct 23, 2026. Spot a mistake? Tell us and we'll correct it in public.
Keep reading
- AI services a local marketing agency can sell (guide)
- 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