- Home
- Skills
- Marketing & SEO
- Salon AI Answer Tracker: ChatGPT, Gemini & Perplexity
Works with the AI Agents you already use
Salon AI Answer Tracker: ChatGPT, Gemini & Perplexity
A monthly log, not a one-off score: records whether ChatGPT, Gemini, AI Mode and Perplexity name your salon. Pairs with Booking Leak Audit.
$7
Salon AI Answer Tracker: ChatGPT, Gemini & Perplexity
Example session with this skill installed
[FICTIONAL test business, written for this demo. Not a real studio. The filled log (ex2-lash-score.log-2026-10.csv) is SAMPLE DATA invented to show the scoring method: no row is a real answer from ChatGPT, Google, Gemini or Perplexity, and every row is labelled fictional/sample. Competitor names are made up. Phone numbers use the 555-01xx fictional range; links use the reserved .example domain.]
User message
Score this round for Kindred Lash Studio, Columbus OH (Clintonville). Main service: volume lashes. Specialty: wispy hybrid sets.
Our facts: Tue-Sat 10-7, closed Sun-Mon. Phone (614) 555-0175. Address 88 Sample Ln, Columbus OH. Booking: kindredlash.example/book.
We take cards and Apple Pay, not cash only. Our old phone was (614) 555-0122. Yelp page exists but I never claimed it. No Bing Places that I know of.
Log attached: ex2-lash-score.log-2026-10.csv
- Read your context and instructions
- Compiled the salon ai answer
Salon AI Answer Tracker - Kindred Lash Studio, Columbus, OH, round 2026-10
Evidence: 24 rows, all labelled fictional/sample (demo data, not real assistant answers). Protocol held: one tester, one MacBook, Columbus OH, all on 2026-10-14; ChatGPT, AI Mode and Perplexity signed out, Gemini signed in with Activity on (as the protocol requires).
Snapshot: In this sample round, Google's assistants named Kindred and got its facts right; ChatGPT and Perplexity mostly didn't, and gave the old phone and wrong hours from Yelp. Nobody named Kindred for wispy hybrid sets.
| Assistant | Discovery named (of Q1-Q4) | Best position | Q5 facts right | Booking link given | Read |
|---|---|---|---|---|---|
| Google AI Mode | 3/4 | 1 (Q4) | 4/4 | yes (Q2, Q5) | named consistently this round |
| Gemini | 2/4 | 2 (Q4) | 3/4 (link missing) | no | named sometimes |
| Perplexity | 1/4 | 3 (Q4) | 1/4 (hours, phone wrong; link missing) | no | named sometimes |
| ChatGPT | 0/4 | - | 1/4 (hours, phone wrong; link missing) | no | not named this round |
What the answers show
Google vs. the rest: AI Mode and Gemini cited the Google profile or Maps and named Kindred on Q1 and Q4. In these rows ChatGPT and Perplexity cited yelp.com every time and named Kindred once (Perplexity Q4).
Same wrong facts twice: ChatGPT Q5 and Perplexity Q5 both gave "Mon-Fri 9-5" and your old phone (614) 555-0122, citing only yelp.com. Your unclaimed Yelp page probably holds both.
- Specialty is invisible: Q3 ("wispy hybrid sets") named Featherline Lashes on all four and Kindred on none.
- Booking link: only AI Mode gave one.
- False claim: Perplexity Q6 said "cash only", citing a 2022 Yelp review. You say you take cards and Apple Pay.
- Competitor source: booksy.com was cited for competitors on ChatGPT Q2 and Perplexity Q2.
Check this: AI Mode Q6 says "deposits are non-refundable". You didn't give me your deposit policy, so check that it matches.
Sources cited most
- yelp.com: 12 rows (you can claim)
- google.com, Business Profile or Maps: 12 rows (you can claim)
- columbuslashguide.example: 2 rows (editorial)
- booksy.com: 2 rows (other: a booking marketplace; you'd need a profile there)
- kindredlash.example, your site: 2 rows (you control)
Fixes, in order
**F3 Wrong fact (hours, phone on ChatGPT and Perplexity)** -> claim the Yelp page (Yelp for Business), set Tue-Sat 10-7 and (614) 555-0175, then search the web for "555-0122" to catch other old listings. 30-60 min.
2.
**F2 Not named on ChatGPT** -> on the claimed Yelp page, set the lash extensions category, services and booking link (Yelp is the BrightLocal study's top ChatGPT source). Check robots.txt doesn't block `OAI-SearchBot`. 1-2 h. Perplexity has no published source study, but its rows here cited yelp.com too. No guaranteed mention.
3.
**F5 Specialty missing** -> add a short website section, "Wispy hybrid sets in Clintonville" (what, who it suits, price), and one line in the Google and Yelp descriptions. 1 h.
- F4 No booking link -> show kindredlash.example/book as a text link on the home page and add it to Yelp. 30 min.
F7 False "cash only" -> once Yelp is claimed, reply to the 2022 review ("We take cards and Apple Pay") and add payment types. Don't ask for Yelp reviews; Yelp doesn't allow it. 15 min.
6. F8 Booksy cited for competitors -> optional; only if you'd take bookings there.
Next round
- 2026-11, same week, same laptop and sign-in states, from Columbus, same 24 questions.
- Add Q5 on Apple Maps: its place cards often show Yelp data, so the old phone may be there too.
- In
notes, list which fixes were done and when.
Not checked
- Whether the Yelp page really shows the old phone and hours (inferred; confirm)
- Bing Places and Apple Maps listings
- Your website's robots.txt
- None of the 24 rows is a real answer. Run the protocol live before acting on any of this for a real studio.
Connects securely to your tools. The creator never sees your data.
About this skill
Monthly tracker; pairs with Booking Leak Audit. It doesn't audit your booking funnel. It records, the same way each month, whether AI assistants name your salon, barbershop, tattoo, lash or brow studio in the US, and whether they get your facts right.
Plan mode fills in six fixed questions for your business and city: best places, book online, your specialty, your neighborhood, your hours/phone/booking link, and what customers say. It also sets the rules that keep rounds comparable: a new chat each time, the same device, and no "near me". ChatGPT, AI Mode and Perplexity are run signed out where they allow it. Gemini is run signed in with Activity on, because Google says its Maps answers need that. You also get a ready CSV log with 24 rows (4 assistants x 6 questions), and every column is defined.
Score mode reads the filled log. It gives a mention rate per assistant, accuracy for hours, phone, address and booking link, and the sources the answers cited (you control / you can claim / editorial). Each finding is tied to a fix: Google profile, Yelp, the Apple Maps card, website text, robots.txt for OpenAI's search crawler, booking-link placement. Compare mode checks that the protocol was held before calling anything a change.
Built for salon owners and the people who market them: salon-specific questions, about 30-45 minutes a round. You run the questions in the assistants yourself. Every answer is labelled observed, reported by user, or fictional/sample, and unlabelled rows are left out of the scores. It never makes up what an assistant said, and it doesn't promise rankings or bookings.
Not affiliated with OpenAI, Google, Perplexity, Microsoft, Apple or Yelp.
US-based salons, barbers, and studios only. Protocol requires 30-45 minutes of manual logging. Does not audit booking funnel logic.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 7 days ago
Permissions
No special permissions declared or detected