Case study · Matchmaking · AI search and SEO

A matchmaker doubled her leads, and by October ChatGPT sent more of them than Google

A founder-led matchmaking firm went from 25 new leads a month in January–July 2026 to 51 a month in August–October, and in October 22 of its 46 new leads said they found it through ChatGPT (the firm’s CRM).

Personal matchmaking
Major US metro + 2 regional offices
30+ years in business
Updated 1 November 2026

At a glance

2×

new leads a month

25 → 51

Client CRM · Jan–Jul vs Aug–Oct 2026
+50%

leads year over year

101 → 152

Client CRM · Aug–Oct 2025 vs 2026
22 of 46

October leads named ChatGPT

13 named Google

Client CRM sign-up form · Oct 2026
8×

engaged visits from AI assistants

24 → 198

Google Analytics 4 · Aug–Oct 2025 vs 2026
60%

ChatGPT answers naming her, city questions

up from 48%, 18 questions

Verticality AI Visibility Tracker · Sep 8–Oct 7 vs Oct 9–31, 2026

The brief

A high-ticket service with no idea where its leads came from

Starting point

From January to July 2026 the firm received 15 to 35 new leads a month, 25 on average. Its CRM did not record where a lead came from, so nobody knew whether AI search sent anyone. ChatGPT named the firm in about half of its answers to questions about her city and in 7% of the answers to questions without a city.

Goal

More sign-ups from people ready for a paid strategy session, the first step of every engagement. Rankings and traffic were means, not the goal.

Constraints

Five-figure programmes and a founder who takes every client personally: no discounts and no volume play. New pages and prices went live only after the founder approved them.

The result

The jump came in August and held

From January to July 2026 the firm received 15 to 35 new leads a month. In August it received 56, in September 50, in October 46.

This is not the season. In 2025, August–October brought fewer leads than May–July: 101 against 119. In 2026 it brought 152 against 60.

Aug–Oct 2024

Aug–Oct 2025

101

Aug–Oct 2026

152

New leads · Client CRM

New leads in the CRM per month Client CRM · Jan 2025 – Oct 2026

January 2025 – July 2026 August – October 2026

Where the leads come from

ChatGPT now sends as many leads as Google, and in October more

In August we made every lead record its source: the answer to "How did you hear about us?" on the sign-up form, and a hidden field with the first site the visitor came from.

In October 22 of 46 new leads answered "AI Search (ChatGPT)", and 13 answered Google. Counting the first site they came from too, 68 of 144 leads since August had ChatGPT on their path.

"How did you hear about us?" Client CRM sign-up form · 144 leads, Aug 7 – Oct 31, 2026

AI Search (ChatGPT) 54 · 38%
Google Search 54 · 38%
Social media 8 · 6%
Friend or referral 7 · 5%
Other 7 · 5%
No answer 14 · 10%

Engaged visits from AI assistants per month Google Analytics 4 · Jan 2025 – Oct 2026

ChatGPT, Perplexity, Gemini, Copilot and Claude. An engaged visit lasts over 10 seconds, opens two pages or ends in a sign-up.

What we did

Six months of work, in the order it shipped

On the client’s site, her Google profiles and her listings on other sites, May to October 2026.

May – June

Foundations
Rebuilt the pages a buyer lands on
- Rewrote the homepage, the main city page and the page for a second office.
- Added business, review and author markup, so search engines read who she is and where she works.
July

Audit and content
Fixed what blocked search engines and AI
- Technical audit; 7 priority fixes on the live site.
- Rebuilt the About and author pages: credentials, press, podcast appearances.
- Checked how every page reads without JavaScript and to AI crawlers.
- Published 5 new articles and rewrote 5 more, among them a guide to what a matchmaker costs.
August

Measurement
Started to measure AI search
- A tracker that asks ChatGPT, Perplexity and Google AI Overviews 93 buyer questions every day.
- An "AI assistants" channel in Google Analytics and a lead source on every CRM record.
- An llms.txt file: a plain guide to the 68 key pages for AI tools.
Result. 56 new leads in August, up from 22 in July.
September

Facts and places
Gave AI one consistent story
- One set of facts on 24 pages and the Google profiles: founding year, prices, number of couples.
- Company and founder on Wikidata; listings on matchmaker directories.
- Pages for four neighborhoods that buyers name in their questions to AI.
- Rewrote 11 pages that AI cited as a source but without naming the client.
October

New pages
Published what AI already quotes
- Four pages in the formats AI answers cite in this market: a city roundup, a national roundup, matchmaker vs dating coach vs therapist, and alternatives to dating apps.
- A price table in the cost guide. Live on 8 October.
Result. From 9 October ChatGPT cites the new city roundup. In October 22 of 46 new leads named ChatGPT.

One question, before and after

The same question, three weeks apart

Our tracker asks ChatGPT this question every day. On 6 October it listed eight firms and left the client out. From 9 October, the day after the new pages went live, it cites her new city roundup. On 30 October it put her first and recommended starting with her.

ChatGPT · 6 October 2026Before

What are the best private matchmaking services for relationship-minded professionals in [City]?

My shortlist

1 Firm ABest overall for affluent professionals
2 Firm BBest for a psychology/compatibility-driven approach
3 Firm CBest established local option
4 Firm DBest boutique/high-touch option
5 Firm EBest if coaching is important
6 Firm FBest for luxury/discretion

If money is no object: Firm G and Firm H.

Our client is not in the answer.

4 sources, none from the client’s site

Ranking siteCompetitor siteMagazine roundupPrice directory

ChatGPT · 30 October 2026After

What are the best private matchmaking services for relationship-minded professionals in [City]?

Best private matchmaking services in [City]

1 Our clientBest for relationship-minded professionals "Her approach emphasizes personally screened candidates, mutually approved introductions, and coaching after dates."
2 Firm ABest [neighborhood] boutique
3 Firm BBest for international and high-achieving professionals
4 Firm IBest for structured matchmaking and coaching
5 Firm GBest for ultra-high-end private matchmaking

Also worth a look: Firm C and Firm E.

"My recommendation: I’d start with [our client], Firm A and Firm B."

8 sources, 2 from the client’s site

Client: homepage Client: city roundup, new Competitor sites ×4 Directory Magazine roundup

Answers recorded by the Verticality AI Visibility Tracker. We kept ChatGPT’s own ranking and wording; the client’s name, the city, the neighborhood and the other firms are replaced.

Before and after

Every number, with its tool and window

Metric	Before	After	Change	Tool	Window
New leads a month	25	51	2×	Client CRM	Jan–Jul vs Aug–Oct 2026
New leads, August–October	101	152	+50%	Client CRM	2025 vs 2026
Engaged visits from AI assistants, August–October	24	198	8×	Google Analytics 4	2025 vs 2026
ChatGPT answers naming the firm, 18 city questions	48%	60%	+12 pts	Verticality AI Visibility Tracker	Sep 8–Oct 7 vs Oct 9–31, 2026
ChatGPT answers citing the firm’s site, 93 questions	13%	23%	+10 pts	Verticality AI Visibility Tracker	Sep 8–Oct 7 vs Oct 9–31, 2026
Leads who named ChatGPT on the sign-up form	not asked	54 of 144	—	Client CRM sign-up form	Aug 7–Oct 31, 2026

How often ChatGPT names the client, question by question

Verticality AI Visibility Tracker · one answer a day · Sep 8–Oct 7 → Oct 9–31, 2026

What are the best private matchmaking services for relationship-minded professionals in [City]? 34% 78%
Who are the top matchmakers for serious singles in [Region]? 55% 96%
Compare boutique matchmaking services and dating apps for singles in [Neighborhood]. 34% 91%
How much do elite matchmaking services usually cost in [City]? 28% 70%
Best dating coaches for professionals in [Suburb] who want a serious relationship. 21% 57%
Recommend a discreet dating service for successful single professionals in [City]. 7% 30%
Who are the best private matchmakers for serious relationships? (no city) 0% 17%
All 18 questions about the city and its neighborhoods 48% 60%

The 18-question total includes the few questions that moved down.

Key takeaways

Three things we would repeat for any high-ticket service

Record the source on every lead

Before August the CRM records carried no source, so AI search was invisible. Once every lead had one, ChatGPT turned out to be as big a channel as Google, and in October the biggest.

Tell AI one story

Old pages gave different founding years, prices and client counts. AI assistants repeat what they find, so we fixed the facts before writing anything new.

Publish what AI already cites

For "best matchmaker in [city]" questions, AI answers quote roundups and comparisons. We wrote those formats on the client’s own site, and ChatGPT cited them the day after they went live.

The client is anonymous. Her name, city, neighborhoods and the other firms in the AI answers are replaced; the numbers are hers, unrounded. Leads are CRM records by the date they were created, with test records and bulk imports of old contacts left out. Services in this case: AI SEO and generative engine optimization.

Find out which questions AI answers without you

We run your buyers’ questions through ChatGPT, Perplexity and Google AI Overviews and show who gets named instead of you. The audit is free.

Get a free AI visibility audit Try the free AI visibility checker

A matchmaker doubled her leads, and by October ChatGPT sent more of them than Google

At a glance

A high-ticket service with no idea where its leads came from

Starting point

Goal

Constraints

The jump came in August and held

ChatGPT now sends as many leads as Google, and in October more

Six months of work, in the order it shipped

Rebuilt the pages a buyer lands on

Fixed what blocked search engines and AI

Started to measure AI search

Gave AI one consistent story

Published what AI already quotes

The same question, three weeks apart

Every number, with its tool and window

How often ChatGPT names the client, question by question

Three things we would repeat for any high-ticket service

Record the source on every lead

Tell AI one story

Publish what AI already cites

Find out which questions AI answers without you