How an AI Assistant Decides Whom to Name
An AI assistant does not rank pages. It reads a handful of sources, then writes a short list of businesses those sources agree on. Which sources depends on the platform: Perplexity and Claude lean on directories, ChatGPT and Google AI Overviews lean on the businesses' own websites, Gemini adds social profiles. Getting named means being present and consistent in the two or three sources each platform reads for your category.
The sources are nameable per category. Accounting firms get pulled from Clutch, ThreeBestRated and ClearlyRated. Contractors and plumbers come from BBB, Houzz and HomeStars. Clinics come from provincial portals and clinic finders. In trades and health, review volume separates the businesses named from the ones passed over. In professional services, the separation comes from the website itself: the city in the page title, service pages that answer the question in plain words, and a blog that shows the firm is active. LocalBusiness schema markup, the fix most agencies lead with, did not separate named from unnamed businesses in our September 2026 check across 24 city and category pairs.
The list also moves. In that same check, more than half of the businesses ChatGPT named appeared in only one of four draws of the same question. A single screenshot proves nothing in either direction, which is why the program measures across repeated draws and reports the share of draws you appear in, never a one-time yes or no.
What We Do Each Month
Month one sets the baseline. We write the questions your customers actually ask ("who should I call for a furnace repair in Sherwood Park", "best optometrist near Whyte Avenue"), run each of them repeatedly across ChatGPT, Google AI Overviews, Perplexity and Copilot, and record every business named and every source cited. You receive the list of competitors being recommended instead of you, the sources each platform drew them from, and the gaps on your side: a directory profile you do not have, a service page that never says what city you serve, a review count a third of the businesses named.
The fixes are ordinary and specific. Service pages rewritten to answer the question in the first paragraph. Business name, address, phone and hours made identical across the directories your category is drawn from, with missing profiles created. A review request process that runs every week. Structured data kept clean as hygiene rather than sold as the cure. Each fix is a line item in the monthly report with the date it shipped.
Every month after, the same questions run again. Your share of draws, the competitors named alongside you, and any new source a platform has started citing are reported next to the fixes that shipped, so you can see which change moved the answer.
What the Report Shows
Three lines, kept apart on purpose. Mentions: how often each platform names you across the question set. AI referral visits: sessions that arrive with a ChatGPT, Perplexity or Copilot referrer, the visits that can be identified as coming from an AI answer. Tracked inquiries: the calls and form submissions from those visits, matched in the same tracking that reports your search and advertising. Most AI-driven visits carry no referrer at all, so the visits line understates the channel. The mentions line tells you whether the work is landing; the inquiries line tells you whether it is paying.
What It Costs
AI visibility is not sold as a separate line item, because the fixes are the same pages, listings and reviews that search visibility depends on. It runs inside web optimization from CAD $1,500 a month or local SEO from CAD $800 a month, on a flat fee with no lock-in after the first three months. The free AI visibility audit shows where you stand before you spend anything. Read why we charge flat fees.