How TofuBofu works
From scan to action plan in a few minutes.
You enter your website
Our AI researches your company by crawling your site. It identifies your industry, competitors, service lines, and the queries your buyers ask. You review and adjust before we scan.
We generate high-intent queries
Not generic questions. Decision-stage queries: "best [your service] for [industry]", "top [your category] companies 2026", "best alternative to [your competitor]." These are the queries that drive pipeline.
We query all 6 AI engines
Each query goes to ChatGPT, Claude, Gemini, Perplexity, Google AI Mode, and Microsoft Copilot. We capture the full response, extract which brands get recommended, and verify with server-side string matching. If your name is not in the response, it counts as absent. No inflated scores.
We crawl your website
We check your homepage, services page, about page, and contact page for schema markup (Organization, FAQ, Service, LocalBusiness), robots.txt AI bot rules, and content structure. We compare against what your competitors have.
AI diagnoses the gaps
GPT-4o analyzes the results: your visibility data, competitor data, and site audit. It generates root causes, content gaps, and specific recommendations grounded in the actual scan data, not generic advice.
You get deliverables
Not recommendations. Deliverables. Content briefs with title, outline, word count, target queries, and schema to include. Ready-to-paste JSON-LD schema for your site. A 30-day action plan with 5 items, each referencing which queries it fixes.
Frequently asked questions
How is the TofuBofu AI Visibility Index calculated?
Your TofuBofu AI Visibility Index is the share of buying-intent queries, across all 6 AI engines, where AI actually names you. We focus the scan on decision-stage queries (the ones closest to a purchase) rather than diluting it with generic awareness questions, so the number reflects real pipeline risk, not vanity mentions.
What AI engines do you cover?
All 6: ChatGPT (GPT-4o), Claude (Haiku), Gemini, Perplexity, Google AI Mode, and Microsoft Copilot (the last two via search API).
How do you prevent false positives?
After our AI analyzes each response, we run server-side string verification. If your brand name does not literally appear in the response text, we override the analysis to absent. We built this after discovering GPT-4o-mini was inflating scores by marking brands as mentioned when they were not.
What is a content brief?
A specific, actionable writing plan. Example: Title: IT Support for Dental Practices. Target queries: best IT support for dental clinics, MSP for dental offices. Outline: 5 sections covering compliance, cost, features, comparison, FAQ. Word count: 2,500. Schema to include: FAQPage, Article.
Does the scan touch or change my website?
No. The scan is read-only: it asks the AI engines your buyers' questions and records who gets named, it does not log into your site or change anything. Later, on a paid plan, any content we draft only goes live after you approve it, and we publish through your CMS with your permission. Nothing touches your live site without your say-so.
How is this different from an SEO tool like Ahrefs or Semrush?
SEO tools measure Google: rankings, keywords, backlinks. TofuBofu measures the AI engines, which use different signals, and then builds the content to close the gaps rather than only reporting them. You can rank well on Google and still be invisible in ChatGPT, because they are different systems. This is measure-and-fix for AI answers specifically, not a Google-rank dashboard.
How often does a scan run, and how fast does the number move?
Track runs one scan a month; paid plans scan weekly. How fast the number moves depends on the engine: search-grounded ones like Perplexity and Google AI Mode re-read the live web and can shift within weeks, while ChatGPT and Claude update their training memory on model cycles, which is slower. You watch the trend month over month rather than expecting an overnight jump.
What do I actually get after a scan?
Your TofuBofu AI Visibility Index (the share of buying questions where AI names you), a per-query and per-engine breakdown of exactly where you are missing, the competitors AI names in your place, and a ranked list of the specific gaps to close. On paid plans, each gap maps to a content brief we then write.