Each category is weighted equally at 25%. A brand that nails three and fails the fourth still loses to a brand that does all four well.
For each brand we crawl the live site, parse structured data, check the AI crawler robots rules, pull link and authority data, and audit review-platform and entity-level presence. No paid AI queries are run for the index itself, which keeps the cohort affordable to refresh every month.
Each of the four categories is scored 0 to 100 based on the signals in that category. Individual signals use the same pass/partial/fail system as the paid report.
The four category percentages are averaged (equal 25% weighting each) to produce the final 0 to 100 score. Letter grades map to the same thresholds as the paid audit: A 80+, B 65 to 79, C 50 to 64, D 35 to 49, F below 35.
On the top brands each month we run a small batch of real category queries through ChatGPT, Perplexity, and Google AI Overviews. The result is shown as a callout, not folded into the score. The full mention-rate testing is reserved for the paid audit, where every brand gets it.
Every brand profile is reviewed by a human before publishing. We check for data quality issues, write the one-line verdict, and flag any brand the automated pipeline has mis-categorised.
We are scoring AI readiness, not SEO as a whole. A few signals you might expect are not part of the composite score:
Matters for business, not for AI mentions. A brand can rank #1 on Google and still be uncited by ChatGPT.
Out of scope. Ads do not feed AI answers.
Weak AI signal. We measure what AI cites, not what the brand posts.
Not our call. We measure discoverability, not whether the product is good.
We source from Crunchbase News, TechCrunch, and StrictlyVC each month. We look for startups that raised funding in the previous 60 days with live websites, Seed through Series C, across a mix of sectors. We exclude mega-rounds from established companies and brands already working with Busca.
Real AI mention rate testing means running paid queries through ChatGPT, Perplexity, and Google AI Overviews per brand, every month. That is the most expensive part of our paid audit, and we cannot do it across 20 to 30 brands monthly without breaking the index economics. Readiness signals are what determine whether AI mentions you, so scoring readiness is the cheapest honest proxy. We do spot-check actual AI mentions on the top brands each month as a separate callout.
The data collection is fully automated. Crawl4ai handles the on-site signals, link and authority data is pulled programmatically, and entity-level signals are checked against a fixed list of platforms. The 0 to 100 score is calculated from the same formula every time. The editorial verdicts (one-line summaries on each brand) are written by a human reviewing the audit output.
Yes. If you have made changes since publication and want the score refreshed, email hello@busca.agency and we will re-run the audit. Rescores are reflected on the brand profile page but the published rank for that month does not change retroactively.
Yes, on request. Email hello@busca.agency. We will remove the brand from the current issue and future issues unless and until you opt back in.
AI search tools are how a growing share of buyers research products now. If ChatGPT or Google AI Overviews do not mention you when buyers ask for recommendations in your category, your competitors get the click. Funded startups have the budget and the timeline to fix this before it compounds.
Get the same audit used for the index run on your site. Full report in 24 hours.