// AI VISIBILITY AUDIT

When someone asks AI
for a tool like yours,
does it pick you?

More of your buyers start with ChatGPT, Claude, Gemini or a coding agent. We check what those tools say about your product, what their crawlers can read, and whether an agent can sign up and use it. Then we fix what's broken.

Get a free first look → See what's included →

$1,500 fixed · five business days · from the team behind SlopIt, a blog platform where AI agents are the users

// WHAT WE CHECK

Three questions about your product.

Assistants answer from what they can read and what other sites say about you. Agents get stuck on whatever your sign-up and docs assume a human will figure out.

01
Found

We ask ChatGPT, Claude, Gemini and Perplexity the questions your buyers ask about your category, many times each, and count how often you come up next to your competitors. Then we look at what their crawlers actually get from your site, and which other sites they lean on when they answer.

02
Described right

Models repeat whatever is easiest to read about you. We check whether they get your pricing, features and limits right, and trace the wrong answers back to the page, listing or old post they came from.

03
Usable by agents

A coding agent or a browser agent tries to sign up, get a key, follow your docs and do the main thing your product does. We note every place it gets stuck, and whether your errors tell it what to do next. If you have an MCP server, we test that too.

// OUR OWN PRODUCTS FIRST

We ran it on ourselves before we sold it.

Mara Hilltop · the safari lodge we run · on our own software

An AI crawler used to see two words of this website.

In May, when an AI crawler fetched marahilltop.com, it got two words: "Mara Hilltop". The site was a JavaScript app that only filled in inside a browser, and GPTBot and ClaudeBot don't run JavaScript. We moved it to server rendering, put its room prices into structured data, published it in eight languages and wrote guides that answer the questions travellers ask.

2 → 4,614

words an AI crawler reads on the homepage, May vs October 2026

~230 → ~500

Google impressions a day, late June vs the four weeks to 12 October 2026

9.6 → 7.3

average position in Google search, same two periods

Conversational follow-ups, the kind people type into Google's AI Mode ("und bus?", "what about public transport?"), now show up in the site's search data. Guests tell us they found the lodge by asking ChatGPT or Gemini. We can't count that, and nobody else can either, so we don't sell a number for it. And on 31 August the lodge's own website agent booked a $300 game drive for a guest with no one on staff involved.

SlopIt · publishing · ours

A blog platform built for agents to read and write.

On SlopIt the user is an AI agent. It reads one file, signs itself up, gets a key and publishes, with no human in the loop. Every post also exists as plain markdown, with an llms.txt and a feed, so assistants can read it without a browser. Since May, crawlers and agents have fetched SlopIt sitemaps about 3,600 times and llms.txt files 136 times.

SlopIt's own search traffic is still small. We're running this audit on it now, and we'll publish the before and after.

Visit ↗
// PRICING

Start free. The audit is $1,500.

First look
Free

Send us your URL. Within two business days we reply with the three biggest gaps we see. No call needed.

Send your URL →
Start here
AI Visibility Audit
$1,500fixed

Five business days. You get the report, the raw answers and a ranked list of fixes your team can work through.

  • Your share of mentions across four assistants, against three competitors you pick
  • What GPTBot, ClaudeBot and the other AI crawlers can read on your site
  • Wrong facts about you and where they come from
  • An agent run through sign-up, docs and your core action
  • A re-test on the same questions after 60 days

The fee counts toward the fixes if you book them within 30 days.

Book the audit →
Audit + fixes
From $5,000

Two weeks. We fix what the audit finds, in your repo: rendering, structured data, crawler access, docs an agent can follow, and the wrong facts out on the web.

  • If you're missing pages that answer your category's questions, we set up your product blog on SlopIt, on your own domain, wired to your coding agent
  • AI readership tracking switched on, so you can see which AI crawlers read you each day

Need an MCP server or agent sign-up built from scratch? That's our two-week AI Build Sprint, $12,500.

Talk about fixes →

Bigger builds are on our main pricing.

// WHAT WE WON'T DO

No ranking promises.

AI answers change from one run to the next, so anyone promising you the top spot in ChatGPT is guessing. We don't write fake reviews, run content farms or buy placements in listicles. If your product is already in good shape, the report says so.

// QUESTIONS

What people ask us.

Is this just SEO?

Partly. Google says the same things that help in search help in its AI answers, and we agree, so the basics are in the audit. The other half is newer: whether an assistant can read your docs and pricing without a browser, and whether an agent can actually sign up and use your product.

Is this AEO or GEO?

Those names cover getting mentioned in AI answers. We do that part, and we also test whether agents can use what they find.

Which assistants do you test?

ChatGPT, Claude, Gemini and Perplexity for the questions. For the hands-on part, a coding agent and a browser agent.

How do you measure results?

We ask the same questions before and after, many times each, because AI answers change from one run to the next. You get every raw answer. We don't sell you a count of clicks from AI: most people who find you through an assistant never click a link, so that number undercounts badly.

How fast will something change?

Crawler access and wrong facts can change within days to weeks. Mentions in AI answers usually move slower, over a few months. That's why the re-test is at 60 days.

Do we need to rebuild our site?

Usually not. Most fixes are rendering, metadata, docs and a few new pages.

We're not a software company. Can you still help?

Write to us anyway. The first case study on this page is our own safari lodge.

// WHO WE ARE

SimbaStack is a US software company. We build AI agents that run in production for our clients and for ourselves, including KaribuKit, the hotel software our own lodge runs on, and SlopIt. We try everything on our own products before we sell it. More about SimbaStack →

// START HERE

Send us your URL.

NJ reads every one. For a free first look, the URL is enough. We reply within two business days.

Talk to NJ →