Get ChatGPT to recommend your business.

When a customer asks ChatGPT, Gemini, Perplexity, or Claude who to hire, whether that is a roofer, a dentist, or a lawyer, the answer names a few businesses. CiteMy.ai builds the data those engines read and quote, work most businesses and agencies have never done, and shows you every week whether they name you.

Free audit by our AI business engine. It crawls your site the way AI crawlers do and emails you a score and the three findings that matter most.

We use your email to send the score, findings, and plan, a few short follow-ups about them (one click stops them), and to reply. By submitting you agree to the Terms and Privacy Policy.

Sign up. Not sure what you get? See an example report

Illustration of the shape an AI answer takes when a business is cited. This is not a recorded answer.

Measured every week, in the answers of

  • ChatGPT
  • Gemini
  • Perplexity
  • Claude

Rather watch? See how it works.

How AI assistants pick which businesses to name, and the work we do so they can name yours.

Captions are on screen, so it works with the sound off. The business in it, Example Air & Heat Co., is fictional, and the AI answers are an illustration of what we work toward, not real results. No one can promise what an AI will say, so we measure it every week.

What owners believe about AI search, and what is actually true.

Most of what worked for Google does not carry over on its own. Here are the beliefs we hear most, and what the data says.

  1. Misconception 1

    Rank on Google and the AI will mention you.

    The new reality

    Google ranks pages against a keyword. AI quotes facts it can attribute.

    Why: SEO wins by shipping a thin page for every buying-intent phrase and linking it hard enough to rank. An AI assistant is not picking a link from a list; it is composing an answer, and it only names sources for facts it can quote. A first-page ranking is not citable data. If your page has none, the AI names someone whose page does: a competitor with quotable prices, service areas, warranty terms, and specs.

    9 of 71

    Windows and doors, from our Phoenix panel

    The engines named 71 businesses; Google does not show 9 of them (13%).

    Google snapshot of Nov 2, 2026 (10 searches: Google Places only; the local pack and organic results were not captured); AI answers from Oct 26, 2026 to Nov 22, 2026 (4 weekly runs); 7 businesses that Google shows are not counted because we were still checking whether they belong in this trade while these runs were read, so whether an engine named them is not known. The study

    1.2%

    percent of brand locations recommended by ChatGPT

    ChatGPT recommended just 1.2% of multi-location brands' locations, versus a 35.9% average appearance rate in Google's local 3-Pack.

    SOCi,
  2. Misconception 2

    Good Google reviews are enough.

    The new reality

    Reviews boost a business the engine can already identify. Entity data is what makes you identifiable.

    Why: An AI answer starts with entity resolution: the engine has to prove your website, your Google listing, and your Yelp page are one business. That is a graph problem, solved by matching schema.org @id and sameAs URLs across every JSON-LD block, with an identical name, address, and phone in each. Reviews strengthen a signal the engine already believes; they cannot start one. When the properties do not line up, the reviews attach to an entity the engine cannot merge with yours, and none of them roll into the answer about you.

    86%

    percent of AI citations

    Across 6.8 million AI citations from ChatGPT, Gemini and Perplexity on location-specific queries, 86% came from sources brands control or manage: websites (44%), listings (42%) and reviews/social (8%).

    Yext,

    33%

    percent of test searches citing Yelp

    In BrightLocal's test of local searches across ChatGPT Search, Gemini, Google AI Mode and Perplexity, Yelp was used as a source in 33% of searches, and every platform used directories.

    BrightLocal,
  3. Misconception 3

    SEO covers AI too.

    The new reality

    SEO chases keyword breadth. AI needs data it can quote.

    Why: The SEO playbook is a page for every buyer-intent phrase, thin enough to ship fast, linked hard enough to rank. AI does not weigh those pages against a keyword; it reads them for a fact it can name a source for. A thin page has nothing to quote, so the answer skips it. The work is different: build the data an engine needs to cite you, in the form it reads.

    • OAI-SearchBotOpenAI's OAI-SearchBot crawler is what surfaces websites in ChatGPT search; sites that block it are not shown in ChatGPT search answers. OpenAI, checked
    • Claude-SearchBotAnthropic runs three bots - ClaudeBot (training), Claude-User (user-directed fetches) and Claude-SearchBot (search indexing); blocking Claude-SearchBot may reduce a site's visibility in Claude's search results. Anthropic,
    • PerplexityBotPerplexityBot is the crawler that surfaces and links websites in Perplexity's search results and is not used for model training; Perplexity recommends allowing it to appear in results. Perplexity, checked
  4. Misconception 4

    AI search is years away.

    The new reality

    Your customers are already asking AI who to hire.

    Why: Customers are using AI assistants for the exact question that sends you work: who should I call. It is growing fast, and the data you publish now is what the engines have to work with when they answer.

    45%& growing

    percent of US consumers, up from 6% (2025)

    45% of U.S. consumers used ChatGPT or other generative AI tools for local business recommendations, up from 6% a year earlier, making AI the third most popular source.

    BrightLocal,

    22%

    percent of US homeowners (recent trades hirers)

    22% of U.S. homeowners who recently hired a trades professional now use AI tools such as ChatGPT to research or find recommendations; 83% start their search online.

    Scorpion,

AI is a computer reading your business.

A person sees your photos, your colors, and a big Call button. An AI assistant's crawler fetches your pages' text and code, and its answer can only use what it could read. Computers need information given the way computers read it.

What a person sees

A good page for people. For a machine it holds no fact to quote: no price, no service area, no hours, no date, and nothing that ties it to the same business on Google or Yelp.

What the AI reads

  • index.html, JSON-LDConsistent

    {
      "@id": "https://exampleroofing.example/#business",
      "name": "Example Roofing Company",
      "telephone": "+1-480-555-0147",
      "sameAs": ["https://www.yelp.com/biz/…", "https://www.bbb.org/…"]
    }
  • data/repair-prices.jsonCitable and reliable

    {
      "@type": "Dataset",
      "dateModified": "2026-09-01",
      "creator": { "@id": "https://exampleroofing.example/#business" },
      "contentUrl": "https://exampleroofing.example/data/repair-prices.csv"
    }
  • llms.txtStructured and formatted

    # Example Roofing Company
    
    > Tile, shingle, and foam roof repair in Mesa, Gilbert, Chandler, and Tempe.
    - [Roof repair prices by roof type (CSV)](https://exampleroofing.example/data/repair-prices.csv)
    
  • robots.txtAccessible

    User-agent: *
    Allow: /
    
    Sitemap: https://exampleroofing.example/sitemap.xml
Illustration. Example Roofing Company is a made-up example; the files on the right are the kind we build.
  • Consistent

    The same facts everywhere: name, address, phone, hours, and prices match on your site, Google, Yelp, BBB, and every directory, so the machine can tell it is one business.

    • NAP
    • hours
    • @id
    • sameAs
  • Citable and reliable

    Facts with a source, a date, and an owner: prices from your own jobs, dated and signed by a named person on your team, and datasets with downloads.

    • dateModified
    • author
    • Dataset
    • CSV / JSON
  • Structured and formatted

    In the formats machines parse: schema.org JSON-LD on your pages, an llms.txt map, the facts in the HTML itself rather than loaded by JavaScript, and a current sitemap.

    • JSON-LD
    • llms.txt
    • HTML
    • sitemap.xml
  • Accessible

    Nothing in the way: the AI crawlers allowed in robots.txt, pages that load fast, and your data published where the engines look.

    • robots.txt
    • OAI-SearchBot
    • Claude-SearchBot
    • PerplexityBot

That is how we win: we give the machine what it needs, in the form it reads most easily.

See what we build

New

A doorbell for AI agents.

AI assistants are starting to run errands for people. When one needs a quote, it needs a way to ask. Agent Doorbell gives it one, so it does not have to fight through a web form or move on to the next business.

How Agent Doorbell works

From your real jobs

A video of every finished job.

Each finished job becomes a short narrated video in a plain-spoken owner's voice, built from the job's own records: what was done, the materials chosen, how long it took, the weather that week, and what the customer said. Never a name or an address. It goes on your YouTube channel and the job's project page. It needs your job software connected (Housecall Pro or Jobber).

Job videos on the services page

Works with the software you already use: Housecall Pro, Jobber, and 9,000+ apps through Zapier.

We connect to it because the jobs your team records every day are the proof AI looks for. We turn them into structured content AI can cite. How the connections work

AI assistants became a consumer habit in three years.

This is no longer an experiment. Here is how fast the assistants your customers use have grown, in the words of the companies that run them and the researchers who track them.

US adults who use ChatGPT

2025: 34% (Pew Research Center)34%20252026: 44% (Pew Research Center)44%2026
Pew Research Center, . Pew notes the question's wording changed across years

900 million

weekly active users (global)

ChatGPT has more than 900 million weekly active users, per OpenAI.

OpenAI,

950 million

monthly active users (global)

The Gemini app has 950 million monthly active users, with daily active users tripling in a year.

Google (Alphabet),

1 billion

monthly active users (global)

Google's AI Mode has surpassed 1 billion monthly active users, per Alphabet's Q2 2026 earnings remarks.

Google (Alphabet),

770.7 million& growing

average monthly referral visits from AI platforms, worldwide

AI platforms sent an average of 770.7 million referral visits a month to websites worldwide (June 2025-May 2026), more than double the year before.

Similarweb,
Forecast

25%Gartner forecast in February 2024 that traditional search engine volume would drop 25% by 2026 as AI chatbots and virtual agents take share. A prediction, not a measurement. Gartner,

What this means for your business

The window is open now

Few local competitors have started building data for AI engines. The first businesses in a line of work to become sources the engines trust get named while the field is still empty.

80%

percent of home-services business leaders

80% of home-services business leaders surveyed say they are unsure how to appear in AI-driven search results.

Scorpion,

Data compounds

An llms.txt, structured data, and datasets keep working after they ship, and every month adds to them. A competitor who starts next year starts behind you.

It is the early days of SEO, again

A few local businesses got serious about Google while their competitors were still in the phone book. AI answers are that moment now.

Get your business into the answers before your competitors do.

Free audit

What we do, every month

An AI assistant is a computer, so every piece gives it your facts consistent, citable, structured, and accessible. It starts with work most businesses and agencies have never done: building the data AI engines read and quote, and putting it where they look. Then your search engine and local presence, the content, the upkeep, and the weekly measurement. Our team does all of it for you.

See all 31 deliverables: what each is, how often, and where it shows up

Where we start

Data, not just content

AI engines cite facts they can quote. We turn your business's own facts into structured data and datasets, then put that data where the engines read — and layer a proprietary dataset per city and line of business on top of them, so your numbers sit next to context an AI cannot find anywhere else.

Prices, specs, service areas, timelines, and warranties from your own jobs become an llms.txt, schema.org types on your pages, and datasets with their own downloads, each dated and signed by a named person on your team. Then our proprietary data layer kicks in: public and private datasets that matter in your market, combined with your facts to produce quotable numbers no competitor is publishing. The exact shape is different for every client — that is deliberate, and it is why a generic agency cannot copy the output. Generic AI posts are what most agencies ship. This is the part they have never done.

  • Price ranges by job typeOffer with a priceSpecification
  • Specs and ratings per productProduct with additionalProperty
  • Service areas by city and ZIPareaServed on the business node
  • Warranty termsOffer and additionalProperty
  • Job and appointment timelinesA Dataset with CSV and JSON

exampleroofing.example/llms.txt

# Example Roofing Company

> Tile, shingle, and foam roof repair and replacement in Mesa, Gilbert, Chandler, and Tempe, Arizona. Prices, service areas, and warranty terms below are published as data.

## Services
- [Tile roof underlayment replacement](https://exampleroofing.example/tile-underlayment): price per square foot, how long it takes, the warranty
- [Foam roof recoating](https://exampleroofing.example/foam-recoat): price per square foot and how often to recoat
- [Roof leak repair](https://exampleroofing.example/leak-repair): what a repair costs by roof type

## Data
- [Roof repair prices by roof type (CSV)](https://exampleroofing.example/data/repair-prices.csv): refreshed monthly
- [Service areas by city and ZIP (JSON)](https://exampleroofing.example/data/service-areas.json)

## Answers
- [How long does a tile underlayment replacement take in Mesa?](https://exampleroofing.example/answers/tile-underlayment-time)
- [What does a roof warranty in Arizona cover?](https://exampleroofing.example/answers/roof-warranty)

A plain-text map of the business for AI engines: services, price data, datasets, and answers, each linked. Most sites have none.

Example for Example Roofing Company, a fictional business. Values shortened where marked with an ellipsis.

Then we distribute it

Distribution is the specialized part. Knowing where the engines actually look, and in what format, is the hard part, and our weekly measurement shows us which sources the engines cite.

  • Your own site

    llms.txt and llms-full.txt, XML sitemaps, and IndexNow pings the moment anything changes

  • Open data portals and dataset search

    Datasets with Dataset schema and open downloads, where dataset search can index them

  • Structured directories and data aggregators

    The business listings and data feeds that other sites and engines draw from

  • Knowledge bases

    Wikidata and similar, where your business is eligible

  • Industry databases

    The industry, professional, and product databases for your line of work

Consistent data and info

An engine has to work out that your site, your Google listing, and your Yelp page are one business. We give it one schema.org entity with a stable @id and sameAs links, remove duplicate and conflicting nodes, and make your name, address, and phone match everywhere, with robots.txt letting the AI crawlers in. We also help you claim your ChatGPT business listing and keep it accurate.

  • @id
  • sameAs
  • NAP
  • Wikidata
  • robots.txt
  • ChatGPT listing

Ongoing content, research, and data

Content, research, and data on cadence. Every week: Google Business Profile posts, local rank tracking, panel measurement, and reactions to local news as it hits. Every month: at least four pieces of content answering questions your customers ask, a panel refresh as the questions shift, an entity-graph audit, and dataset refreshes. Every new Google review gets a drafted reply for you to approve. Every quarter: a deep-dive study with new datasets we build for your market. Your data stays current without you chasing it.

  • Weekly GBP posts
  • 4+ pieces a month
  • Review replies
  • Quarterly study
  • News reactions
  • Entity audit

Measure

Every week our AI business engine asks ChatGPT, Gemini, Perplexity, and Claude the questions your customers ask, several times each, and records which businesses each answer names, cites, and recommends, and which sources it cites.

  • Prompt panel
  • Share of voice
  • Citation sources

Example

What we typically find, and what we build

An example audit of a fictional roofer, Example Roofing Company. The business is made up; the findings are typical of what our audits find.

Read the full example report

What we find

  • No data to cite anywhere on the site
  • Content not structured for AI ingestion
  • No llms.txt
  • Prices, service areas, and warranty terms only in prose
  • No datasets in Google Search Console or dataset search
  • Customer questions answered without numbers
  • Business nodes that disagree
  • Hours, areaServed, and geo missing
  • PerplexityBot blocked, and a Crawl-delay for every crawler
  • Sitemap stale
  • Key content rendered only by JavaScript
  • City pages with no job in them
  • No project pages
  • No rebate or energy-savings page
  • No author or dates on posts
  • Nothing published in the last 90 days
  • Google Business Profile never linked from the site
  • Old address and phone on two directories

What we build and fix, in this order

  1. 1llms.txt createdMost sites have none. We write one that maps the services, the price data, the datasets, and the answers, plus an llms-full.txt with the text of each page.
  2. 2Structured data for the business's own factsRepair prices as Offer price ranges, the cities served as areaServed, warranty terms and roofing specs as Product with additionalProperty, on the pages that state them.
  3. 3Customer questions answered with the business's own dataOne page per question, answered with numbers from the business's jobs, marked up with FAQPage, dated, and signed by a named person on staff. The price table becomes a dataset with CSV and JSON downloads.
  4. 4One business entity, consistent everywhereOne business node with a stable @id that every page references, the duplicate and conflicting nodes removed, and the same name, address, and phone on the site and every directory.
  5. 5The other gaps closedPerplexityBot let in and the Crawl-delay dropped, a fresh sitemap, prices and service areas in the HTML, authors and dates on posts, hours and geo added, old listings corrected, and Bing Webmaster Tools and IndexNow set up.

robots.txt, before

User-agent: *
Crawl-delay: 10

User-agent: PerplexityBot
Disallow: /

robots.txt, after

User-agent: *
Allow: /

Sitemap: https://exampleroofing.example/sitemap.xml
Example robots.txt, before and after. Many sites slow or shut out the AI search crawlers without knowing it.

Our AI business engine

Built on the leading AI models, through their APIs. The system around them is ours.

  • The measurement panel

    Your customers' questions, asked of ChatGPT, Gemini, Perplexity, and Claude every week, several times each, with every business named and every source cited pulled out of the answers and counted into share of voice.

  • The auditor

    Crawls your site the way AI crawlers do and checks what they read: llms.txt, structured data, the facts you publish, robots.txt, and your listings.

  • The data pipeline

    Builds your llms.txt, schemas, and datasets, publishes them, sends them to the places engines read, and keeps them current.

What the weekly report tracks

Your share of the four engines' answers, week by week, with each piece of work marked where it went live. What we build keeps working after it ships, so each piece stacks on the last.

Illustration: a share-of-voice line that rises in steps after each milestone: Baseline audit, llms.txt and data schemas, Datasets published, Questions answered with data, Entity and listings cleaned up, Ongoing updates. Not client data, and not a prediction.Share of voiceWeeks1Baseline audit2llms.txt and data schemas3Datasets published4Questions answered with data5Entity and listings cleaned up6Ongoing updates
  1. Baseline audit
  2. llms.txt and data schemas
  3. Datasets published
  4. Questions answered with data
  5. Entity and listings cleaned up
  6. Ongoing updates
Illustration of how the weekly report tracks share of voice. Not client data, and not a prediction.

Datasets live

Each dataset, where it is published and distributed, and when it last refreshed.

Content published

Each page, the question it answers, the numbers it uses, and whose name is on it.

Listings fixed

Each directory and data source corrected, with the date it changed.

Citations tracked

Every answer that names, cites, or recommends you, per engine, per week.

What the weekly report counts. Measured outputs, never a promised rank.

One price. No contract.

The work does not stop at launch. Sites change, plugins update, directories drift, and the engines keep reading, so we keep the data right and fresh every month.

$2,000a month

Month to month. Cancel any time. No contract.

Covers the build and AI Citation Defense, every month you are a client. What we watch every week

Month to month. Cancel any time. No contract. Everything included

  • It compounds

    The llms.txt, structured data, and datasets we build keep working after they ship, and each month adds to them. Ad spend stops when you stop paying.

  • It stays true

    Facts that stay accurate and consistent keep giving the engines a reason to cite you. We keep them accurate and consistent every month; that is what the maintenance is for.

  • It is yours

    One flat monthly price, no contract, no per-post or per-page fees. You own everything we publish.

Everything the price includes (30)
  • Full audit and written plan: We audit what is on your site, what is missing, and what AI crawlers can read, then write the plan for what we will build.
  • llms.txt and llms-full.txt: A plain-text map of your services, prices, data, and answers, written for AI engines and kept current.
  • Structured data for your business facts: One schema.org entity for your business: price ranges as Offers, service areas, warranties, hours, and licenses or credentials.
  • Product and brand pages: For businesses that sell or install brands: one page per product line or brand you carry, with specs, warranties, ratings, and Product schema.
  • First-party datasets: Datasets from your own jobs, published as CSV and JSON with Dataset schema, refreshed on a schedule.
  • A proprietary data layer for your city and line of business: Your facts joined with public market data for your city and line of business, giving the engines numbers they cannot find anywhere else.
  • The facts that set you apart, said the same way everywhere: Optional: you name the few true things that set you apart, each with its proof, and approve the exact wording. We put those words in your structured data, llms.txt, and quick-facts page, suggest your Google Business Profile description, and give you the same text for each listing you control (BBB, Yelp, manufacturer pages, associations, press), then check each one.
  • Quick-facts page: A short, dated page of questions and answers about your business (license, service area, warranty, what sets you apart, and what AI assistants get wrong), with FAQPage schema and a link from llms.txt. You approve it before it goes live.
  • Quarterly deep-dive study: Original research on your market, published on your site and distributed where the engines read.
  • Distribution where AI engines read: Your data sent to sitemaps, llms.txt, IndexNow, open data portals and dataset search, structured directories, and Wikidata where you are eligible. Each listing and dataset page links back to your site, so the backlinks build as the data spreads.
  • Your ChatGPT business listing: ChatGPT now lets a business claim its own listing (OpenAI's Business Hub, launched November 2026). We walk you through the claim, so your own team owns it and adds CiteMy.ai as your agency, then we keep its details accurate and put its numbers in your monthly summary: how often ChatGPT showed your business, the clicks to your website, and what people were asking about. Claiming does not decide where ChatGPT places you; it makes what ChatGPT shows about you accurate.
  • Agent Doorbell: A direct way for AI agents to request a quote from you for the person they work for. The agent reads what you need, sends the details with that person's OK, and the request lands in your inbox or job software, marked as an AI agent. We also publish the small files agents check first to find it: an MCP server card and an API description.
  • Service-area pages: One page per city you serve, built from your real work there: price ranges, photos, and job details.
  • Project pages: One page per completed job, with before-and-after photos where your work has them, what was done, the neighborhood, and the price range.
  • Job videos: Each finished job becomes a short narrated video, in a plain-spoken owner's voice, built from the job's own records: what was done, the materials and options chosen, how long it took, the weather that week, and what the customer said afterward. Never a name or an address. Posted to your YouTube channel and the job's project page, with its transcript. Needs your job software connected (Housecall Pro or Jobber).
  • Manufacturer page and dealer-locator listings: For businesses that sell or install brands: a page listing the brands you carry and your certifications, plus a check of your listing in each manufacturer's dealer locator (present, missing, or wrong). We write the exact profile copy for each portal, you submit it with your own login (we never ask for your password), and we verify it is live.
  • Rebate and incentive pages: Where programs exist for your work: pages on the utility, state, and local rebate and incentive programs in your area, so the questions about them land on you.
  • Google Business Profile management: Weekly posts, photo uploads from your project pages, and your services kept current.
  • Google review replies: A reply drafted for every new Google review, approved by you, then posted.
  • Local search rank tracking: Your map pack and organic search positions for your key searches, recorded every week.
  • The same name, address, and phone everywhere: Your listings matched across Google, Bing, Apple, Yelp, BBB, and the directories for your industry that the engines read.
  • At least 4 pieces of content every month, at a minimum: any mix of articles, seasonal pieces, and 48-hour news reactions, each counting as one.Articles: A question your customers ask the AI engines, answered with your own numbers and your market's data: pricing, products, how-tos, and project case studies.
  • Seasonal pieces: Published ahead of your season (tune-ups before summer, storm prep before monsoon, tax checklists before April), so the engines read them before your customers start asking.
  • 48-hour local news reactions: When local news touches your line of work (a storm, a heat wave, a rate change, a new law or code update), a short, data-backed response from you goes up within two days.
  • Weekly AI share-of-voice report: Your share of the answers from ChatGPT, Gemini, Perplexity, and Claude to your customers' questions, with the sources each answer cites.
  • Competitor citation watch: When a competitor is cited for a question you care about and you are not, we see it in the weekly data and the work addresses it.
  • Monthly panel refresh: The questions we measure keep up with what your customers ask: new ones added, stale ones retired.
  • Analytics setup: AI assistants set up as their own channel in Google Analytics, plus form conversion tracking.
  • Monthly entity-graph audit and maintenance: Your schema re-checked, AI crawler access monitored, and drift fixed before it costs a citation.
  • Monthly update and a 30-minute call: What changed and what comes next, covering leads, rankings, and AI visibility.

What each one is, how often it happens, and where it shows up

Questions owners ask

Can you promise ChatGPT will recommend us?

No, and nobody honestly can: the engines decide what they say. We do the work that gives them reasons to name you, and we measure every week whether they do. Results describe what was measured, not a promise of ranking.

How is this different from SEO?

SEO works on where you appear in Google's list. An AI assistant does not show the list; it answers, and it quotes facts. We build the data it reads (an llms.txt, structured data for your own facts, datasets), distribute it to where the engines look, clean up what confuses them, and measure the answers themselves. CiteMy.ai can sit on top of your SEO agency or replace it.

Is CiteMy.ai software?

No. CiteMy.ai is a done-for-you service: our team does the work every month, on your site, on search engines, and where the AI engines read. The free tools and the client portal are how you see the work; you never have to run anything yourself.

Can I fire my SEO agency and just use CiteMy.ai?

Yes, and you keep the parts of their work that mattered. You still get at least four pieces of content every month, Google Business Profile posts, review replies, weekly local rank tracking, and new backlinks; ours are aimed at how AI assistants answer rather than at where Google ranks. The content is your customers' questions answered with your own numbers, under a named author on your team. The backlinks come from publishing your first-party datasets where other sites and engines cite them: open data portals, dataset search, industry databases, and Wikidata where you are eligible. And because Google reads the same structured data and clean pages, the work does not hurt your rankings; often it helps.

Do you use AI to do this?

Yes. Our team does the work for you with our AI business engine: the system our team built on the leading AI models from OpenAI, Anthropic, and Google, and runs for you. You never have to log in to it. It measures what the engines say about your business every week, audits your site, and builds, publishes, distributes, and keeps your data current. The measurement panel, the auditor, and the pipeline that builds and distributes your data are ours. You approve every change on your site.

Which AI assistants do you measure?

ChatGPT, Gemini, Perplexity, and Claude, every week, with the questions a customer in your market asks.

Do I have to write anything?

No. We draft each piece, and a named person on your team reviews and signs it, because readers and the engines weigh a real author. It uses numbers from your own jobs, so we will ask you for them.

Who do you work with?

Local service businesses anywhere in the US: any business customers hire for a service, from roofers, plumbers, and HVAC companies to dental practices, law firms, accountants, med spas, auto repair shops, and landscapers. Our outreach and public research started with Phoenix home services; when a business in a new industry signs up, we build the questions its customers ask and add that panel to the weekly measurement. One client per metro and line of business: a client holds every city of its metro for its line of business. When a market is taken, it is taken for as long as that client stays, and the next business in line is on the waitlist.

Why do you build data instead of just writing posts?

Because an AI assistant is a computer reading your business, and it answers with facts it can read and check. We give it your facts consistent everywhere, citable (with a source, a date, and an owner), structured (schema.org JSON-LD, llms.txt, clean HTML, sitemaps), and accessible (AI crawlers allowed, pages fast, data where the engines look). Posts alone do none of that.

What is the proprietary data layer you mention?

On top of your own facts, we engineer datasets per city and line of business by combining public and private data that matters in your market with your numbers to produce quotable rows an AI cannot find anywhere else. The exact mix is different for every client — which is why a generic AI agency cannot produce it, and why we do not list the recipe up front. You see every dataset before it ships and you own all of it.

What is upfront and what is ongoing?

Upfront (first 30 days): the audit and plan, your llms.txt and structured data, service-area and product pages (plus manufacturer and rebate pages where they apply), the first datasets we build with your numbers and your market's data, your name, address, and phone matched across directories, analytics set up with AI assistants as their own channel, and your baseline measurement. Every week after that: your AI share-of-voice report, competitor citation watch, Google Business Profile posts, and local rank tracking. Every month: at least four pieces of content (articles, seasonal pieces, and 48-hour reactions to local news), a panel refresh, an entity-graph audit, and an update with a 30-minute call. Every new review gets a drafted reply for you to approve, and every completed job can become a project page with its photos and a short narrated video. Every quarter: a deep-dive study for your market. The weekly watch and the work that answers it are what we call AI Citation Defense. The full list, with how often each happens, is on our services page.

How do I get started?

Two ways. Check if your city is open: pick your city and line of business, and if it is open, claim it, tell us about your business, and pay on Stripe's secure page; we email you a sign-in link to your portal. If it is taken, join its waitlist. Or start with the free audit and see your score first. Either way it is $2,000 a month, month to month. Cancel any time. No contract.

Is there a contract?

No. $2,000 a month, month to month. Cancel any time. No contract.

What happens if I cancel?

Cancel any time in your billing portal; the subscription ends at the close of the paid month. Everything we published stays on your site, and it is yours. The weekly measurement, the monitoring, and the response work of AI Citation Defense stop, and your market (your metro and line of business) opens to the next business on our waitlist, which then gets our full effort in that market. If your spot is still open later, you are welcome back.

What does the free audit cost?

Nothing, and there is no call to book. You enter your website and email, and the score and findings arrive by email.

Built for window and door companies, roofers, and HVAC companies.

See what our AI business engine says about your business.

Start with your website. Our AI business engine crawls it the way AI crawlers do and emails you a score out of 100 and the three findings that matter most.

Sign up

Month to month. Cancel any time. No contract.

Free. We crawl your site the way AI crawlers do and email you the score, the three findings that matter most, and a step-by-step plan. We use your email to send them, a few short follow-ups about them (one click stops them), and to reply. By submitting you agree to the Terms and Privacy Policy.