AI visibility for local marketers

Get your clients into the AI answer.

Customers now ask ChatGPT, Perplexity and Claude who to call. Algotally asks them the same questions week after week, market by market, and shows who gets named, why, what to fix, and what followed each fix.

One email when sign-ups open. Nothing else.

See a sample report

Run your own business? See Starter: see how often AI names it, and what to fix first.

  • 14-day free trial at launch
  • No card needed
  • Monthly, no contract

Illustrative example: names and numbers are made up.

Free, no sign-up. See whether ChatGPT, Perplexity and Claude name a business when customers ask for a recommendation, in the markets we measure publicly.

3 AI enginesChatGPT, Perplexity and Claude, web search on
Up to 72 answersA week per market since Oct 8, 2026, from real buyer questions. Each market page shows its count.
95% rangeOn every rate, with the answer count behind it
70+ toolsREST API and MCP server on every plan
The shift

Local search has a new front page, and it names three businesses.

Your clients are already asking why ChatGPT recommends the competitor down the street. Most marketers have no evidence-based answer yet.

Customers ask, AI answers

Instead of ten blue links, people get a short list of names. Being on that list is the new top-three ranking.

AI repeats what it reads

Answers come from the directories, reviews and listings AI searches first. Wrong hours, a closed location or a missing listing travel straight into the answer.

The fix is mostly listings, not content

Local AI picks follow listings and reviews. Knowing which sites AI read for a market tells you exactly where your client has to show up.

How it works

Measure. Explain. Fix. Prove.

One loop, from “does AI name my client?” to “here is what changed after the fix”.

Measure

We ask like a customer, week after week.

Buyer questions for each industry and city, answered by AI engines with live web search on. We record who each answer names, in what order, and which pages it cites.

  • Mention rate and share of voice for every business in the market
  • A 95% range and the number of answers behind every figure
  • The exact questions asked, so nothing is a black box
Share of voice · Plumbing, Houston (example)
Bayou Pipe Pros
21%
Gulf Coast Rooter
16%
Lone Star Drain
10%
Acme Plumbing
7%
Northside Flow
5%
Explain

See why AI chose someone else.

Before answering, AI searches the web. We rank the sites it reads in each market, show which ones AI already cites alongside your client, and check whether what AI says about them is true.

  • The directories, review sites and articles AI cites most, per market
  • Which of them AI cites alongside your client, and which it doesn’t
  • Wrong facts AI repeats: hours, phone, services, closures
Sites AI read before answering (example)
yelp.comCited alongside
expertise.comNot checked
bbb.orgCited alongside
Local news “best plumbers” listNot checked
Fix

Get the fixes written for you.

Every report comes with a kit built from its own evidence and your client’s real details, ready to paste or send. Tools tell you what; this hands you the work.

  • Listing kit: how to claim each site AI doesn’t cite it on, with copy to paste
  • Review request kit for the review sites AI reads in the market
  • Outreach, pitch and correction drafts you send yourself, with a call guide
  • Website files: schema markup, a pricing and an FAQ section
  • An action plan in the order most likely to move the number

What the evidence says moves AI →

Action plan (example)
Claim the expertise.com listingDone
Ask recent customers for reviewsDone
Correct Sunday hours on two directoriesTo do
Pitch the local “best plumbers” listTo do
Prove

Most AI advice is unproven. Your results won’t be.

A 2026 review of 45 studies found no technique with a proven, lasting effect on AI recommendations across engines. So nothing here asks you to take a fix on faith: mark a step done and we compare AI answers before and after. A change is only called a win when it clears a 95% noise test, so you never claim luck as results.

  • Before-and-after for each change, with a noise test (Growth)
  • The leads that told the business AI sent them, logged month by month beside the trend, as context rather than proof (Growth)
  • Which kinds of change have moved AI mentions, pooled across all subscribers, with plans ordered by them (Agency)
  • Share links for clients, under your name on Growth and your brand on Agency

What the evidence says moves AI →

Outcomes (example)
4 weeks before7%of answers
4 weeks after19%of answers
Real change. After a listing was claimed, the rise clears the 95% noise test across 380 answers.
Within noise. Another change went from 9% to 10%. Too small to call, and the report says so.
Why Algotally

Other tools tell you if AI mentions your client.
We tell you how sure, why, and what to do.

A typical AI trackerAlgotally
The numberOne run, one scoreRepeated sampling with a 95% range and the answer count
What is askedPrompts you write for your clientThe whole market, with every question shown
The whyGeneric content and “GEO” adviceThe exact sites AI read, and which it cites your client on
The workA report to readListing kits, website files and drafts, written for you
The proofA chart that wigglesBefore-and-after with a noise test
What you can do with it

Built for the jobs local marketers actually have.

Win the pitch

Run the free check on a prospect, show them the competitor AI names instead, and hand over a share link that does the selling.

Free check · share links

Explain a drop

When a client asks why ChatGPT stopped naming them, show which sites changed and who took their place.

Market report · history

Correct wrong facts

Catch AI repeating old hours, a wrong phone number or a closed location. On Growth, get a correction kit to fix it at the source.

Fact check · every plan

Prove the retainer

Before-and-after for each change, tested against noise, so your monthly report shows results you can stand behind.

Outcome tracking · Growth

Report under your brand

White-label share links, a client roster with bulk reports and weekly monitoring, and CSV or JSON export.

Agency

Work from Claude or Cursor

Add the MCP server with one config snippet and run reports, market reports and fix tools from the assistant you already use.

Every plan
Pricing

One client’s retainer covers it.

Flat monthly prices in US dollars. Sign-ups open soon. At launch, every plan starts with a 14-day free trial: 3 reports with the plan’s own features (Agency’s on an Agency trial, apart from team seats, which open when the plan is paid) plus Growth’s, and no card.

  • 14-day free trial at launch
  • No card to start
  • Flat monthly price
  • No contract

Starter

Where a business stands, and how that has moved week by week.

$49/ month

$49 a market, about $3.27 a report · Add one market: $25/month

1market
15reports / mo
1seat
Coming soon
  • Week-by-week history, in reports and share links
  • Fact sheet & hallucination check, checked weekly
  • Which AI engines name the business (ChatGPT, Perplexity, Claude)
  • Business report (snapshot, sites, website check, market size)
  • Market report
  • This week card: each business’s latest week and trend, free

+ 13 more on the pricing page

Pick it if: you own a business and want to know what AI says about it, checked weekly, or you are trying AI visibility with a client or two in one market.

Most popular

Growth

Proof of what worked for the businesses you work on, and how each stacks up against the competitors you name.

$99/ month

$33 a market, about $2.48 a report · Add a market: $35/month · White-label: $49/month · +2 nearby cities per market (up to 6 in all)

3markets
40reports / mo
1seat
Coming soon
  • Everything in Starter
  • Outcome tracking to prove what worked (before and after each change, results on client links)
  • "Prepared by" your name on share links
  • Pitch drafts for every editorial outlet
  • Correction kit for what AI gets wrong
  • Deep sampling (2× answers, tighter ranges)

+ 7 more on the pricing page

Pick it if: you sell AI visibility as a monthly service and need to prove it works.

Agency

Close to everything we see, for a book of clients.

$299/ month

$29.90 a market, about $1.20 a report · +5 nearby cities per market (up to 15 in all)

10markets
250reports / mo
5seats + viewers
Coming soon
  • Everything in Growth
  • Answer evidence and the Engines page (the AI answers that named the business)
  • CSV / JSON export
  • Full white-label share links, plan links and the monthly client report
  • Client roster with prospects, bulk reports and outreach, monitoring, seats, and a quarterly review per client
  • "What works" insights (builds as results are logged: 5 per fix), and plans ordered by what works

+ 6 more on the pricing page

Pick it if: you run a book of clients across markets, or report under your own brand.

No credits. No per-question pricing. Every AI engine on every plan. Opening a saved report and using the kits doesn’t use a report.

Prices in USD per month. Report allowances reset each month. Compare every feature · Why no Google AI or Copilot?

What counts as a market, a report and an answer?
Market
One type of business in one city, e.g. plumbers in Houston, TX. Plans are sized in markets: every report runs inside one of yours, and you can swap one out each month.
Nearby city
On Growth and Agency, a market can add nearby cities in the same state within about 40 miles (a service area), e.g. Katy and Sugar Land for plumbers in Houston. Each is measured as its own market, at up to 36 answers a week, without using a market slot, and business reports show the rate city by city.
Report
One run of a business report (how often AI names a business, where it ranks, the sites AI reads and which it’s missing from, plus the tools to fix it) or a market report (who AI recommends in a market). Reopening a past report is free.
Answer
One reply from an AI assistant (ChatGPT, Perplexity or Claude, web search on) to a buyer’s question. Every rate says how many answers it rests on.
Our promise

No guarantees. Just evidence.

Nobody can promise that AI will recommend a business, and anyone who does is guessing. We show you what AI actually says, how sure that is, and what changed after the work.

Read the methodology → · What moves AI →

Every rate shows its workThe answer count and the 95% range on every rate, and the questions asked, on the report and the market page.
We never contact your clients’ businessesOutreach, pitches and corrections are drafts you send yourself. Nothing is published for you.
Wins are tested, not claimedA change is only called up or down when it clears a noise test.
FAQ

Questions people ask

What does Algotally measure?

Which local businesses AI assistants recommend, and which websites they read before answering. Every week we ask Perplexity, ChatGPT and Claude, with web search on, the questions real customers ask, many times over ("Which plumbers in Houston should I call?"), and record who each answer names and which pages it cites.

Who is it for?

Freelancers and agencies who do local marketing for home service and other local businesses. It shows where a client stands in AI answers, what to fix, and whether the fixes worked.

What is share of voice?

A business’s slice of every recommendation AI made in its market. If answers named businesses 200 times and 20 of those were one business, its share of voice is 10%. We report it next to the mention rate: the share of answers that named the business at all.

How accurate are the numbers?

AI answers vary from one run to the next, so every rate comes with how many answers it rests on and a 95% range, and a change over recent weeks is only called up or down when it clears a noise test. The methodology page explains each step, and our noise study shows how much answers move in our own data. Methodology · Noise study

How much does it cost?

Starter is $49 a month, Growth $99 and Agency $299. Starter can add one market for $25 a month; Growth can add markets for $35 a month each and white-label for $49. Sign-ups open soon; at launch every account starts with a 14-day free trial with 3 reports, the plan’s markets (up to 3) and no card. No credits and no per-question pricing: every AI engine on every plan, and opening a saved report or using the kits doesn’t use a report.

Why not Google AI Overviews, Gemini or Copilot?

Google’s terms don’t allow storing or analysing its AI answers, and we don’t scrape Google, so we measure ChatGPT, Claude and Perplexity through their official APIs. Copilot isn’t covered yet. Listing, review and fact fixes help on Google too, and Google Search Console and Bing Webmaster Tools show your own site’s numbers in their AI features for free. Algotally adds what those don’t: competitors, the whole market, and what AI says about you.

What do I get beyond the numbers?

Tools that write the fixes: the directories AI reads that a business isn’t confirmed on and how to claim them, a review request kit for the review sites AI reads, pitch drafts for the outlets and lists AI cites, website files (schema markup, a pricing section, an FAQ section and a booking link built from the business’s own details), outreach drafts with a call guide, and an action plan whose steps are measured before and after.

Does Algotally contact the businesses I look up?

No. Algotally never contacts a business. Outreach, pitches and corrections are drafts you send yourself, and it never publishes anything to a business’s website.

Can an AI assistant or my own software use Algotally?

Yes. The free check is an open API (GET /api/v1/check), and every account can create API keys for the REST API and an MCP server with over 70 tools, so an AI assistant can run reports and tools directly.

Which markets are covered?

Any type of business in any US city. Each plan holds a set number of markets (Starter 1, Growth 3, Agency 10), and you choose them yourself in the app: one we already measure is ready at once, a new one starts being measured the same day. Agency can also ask us to measure markets beyond its own 10. Published markets are listed on the markets page, each with the sites AI cites most there.

Do you guarantee AI will recommend my client?

No, and be wary of anyone who does: AI answers change from run to run and no tool controls them. Algotally shows what AI actually says, which sites it relied on, what to fix, and whether the numbers moved after the work, tested against noise.

Is there a contract?

No. Plans are monthly. Billing isn’t live yet; once it is, you cancel from your account and the plan runs to the end of the month already paid for. Until then, write to support@algotally.com. A canceled account stays readable, so past reports and history are still there.

Is there a free way to try it?

Yes. The free check at algotally.com/check tells you whether AI names a business in its market, and which sites AI reads there, with no account. The sample report shows what a full report adds. For developers, the website audit API (GET /api/v1/audit?url=) scores how readable a site is to AI, also free.

For AI assistants and developers

Everything a person can do here, software can do too.

No account needed for the free check and the website audit. Accounts get API keys for the REST API and the MCP server at /mcp. Setup for Claude, Cursor and ChatGPT. A summary for AI readers is at /llms.txt.

GET /api/v1/check?name=&industry=&city=&state=
  → does AI name this business?
    (markets: GET /api/v1/markets)

GET /api/v1/audit?url=
  → how readable is this site to AI? (0–100)

POST /mcp   Authorization: Bearer si_live_…
  → over 70 tools: reports, markets, fixes, outcomes

Find out who AI recommends
in your client’s market.

Sign-ups open soon, starting with 3 free reports. Leave your email and we’ll tell you when.

One email when sign-ups open. Nothing else.

Run a free check now →

  • No card needed
  • Your plan’s features, plus Growth’s, in the trial
  • Plans from $49 a month