Marketing · Attract module · search & AI answers
AI assistants don’t mention us
A buyer asks an AI assistant who does what you do, and you aren’t named. What makes an assistant name a company, and how to check and fix it.
Starts on AI visibility · $95 a month
In your words
A customer told us he’d asked ChatGPT for “commercial refrigeration repair near Laval” before he called us. He got four names. Ours wasn’t one of them, and we’ve been doing this for twenty-two years.
We tried it ourselves with the questions our buyers actually ask. Same thing: the assistants named two competitors, a directory and a company that closed last spring. Nobody here knows what the assistant read, or what we would even change.
Why it happens
An assistant with web search names the companies it can read and verify. It looks for a plain statement of who you are, what you do and where, in text it can parse without running your site’s JavaScript, and it leans on pages that other sources point to. A homepage that says “Excellence since 2003” over a photo carousel gives it nothing to quote.
Often the door is shut before the reading starts: a robots.txt written years ago that blocks every bot it doesn’t know, which now includes the crawlers assistants use. Or the services live in a PDF brochure, or only in a dropdown menu. The assistant isn’t ignoring you; it simply has more to go on for the companies it names.
What it costs
Buyers who start with an assistant arrive with a short list already made. If you are not on it, you are not in the comparison, and you never see the inquiry you lost: there is no missed call, no form, nothing to count.
It also compounds. Assistants cite the pages they trust, those pages get read more, and the companies named today are easier to name tomorrow. Distributors fared worst on this in our own front-office report: 74 out of 100 on AI answers, from automated checks of their homepages.
How Humonix fixes it
01Agent
Check what an assistant can read
The free check reads your public homepage, robots.txt, sitemap.xml and llms.txt: 34 named checks, three scores, in about a minute, with up to three competitors beside you.
02Agent
Write down what buyers ask
With AI visibility you keep the questions your buyers really ask, checked every month (every week on Hue Pro, in English and French). Each is put to the connected AI assistants with web search, and every run records whether you were named, where, and which sources were cited.
03You
Approve the fixes
The Attract module drafts the fixes from your own facts: page title, description, business schema, FAQ, llms.txt. You approve or decline each one; today a person places them on the page.
04Agent
Watch the trend
Scores by month and answers by question, so you can see what changed after each fix instead of guessing.
01Agent
The department
The Attract module does the reading, the asking and the drafting. It never publishes to your site by itself.
02Hue
Hue’s part
Hue answers on your site with the same facts the fixes are written from, so what an assistant reads and what a visitor is told match.
03You
Your yes
Every fix waits for your yes in Approvals, and the decision is on the ledger. No one can promise a mention; a score is a snapshot of your homepage, not a ranking promise.
What you see
Approvals: each drafted fix (a title, a description, your business schema) waits for a yes, with what it was written from.
Where to start
AI visibility
Read-only, nothing to install: your buyers’ questions checked on AI engines every month, against three competitors, every fix ranked. Month to month.
Questions
Can you get us into ChatGPT’s answers?
No one can promise that. We measure whether assistants can read you and whether they name you, and we fix what you approve.
Which assistants are asked?
ChatGPT, Perplexity, Gemini, Claude and Grok, each through its official API with web search, once it is connected for your workspace; the app shows which are. Google AI Overviews are not checked yet.
What does the free check read?
Your public homepage, robots.txt, sitemap.xml and llms.txt. Public pages only, and our crawler respects robots.txt.
Do we need a new website?
Usually not. Most fixes are words and structured data on the pages you have. When a page is missing, the Marketing department can put one in front of your current site.