# Do the companies on these lists say AI on their homepage?

Homepage text of every company on four lists, scored for a visible mention of
AI. Crawled 2026-09-04 to 2026-09-04. Homepage only, one page per company.

## Files

- `fortune500.csv` - one row per company on the 2026 Fortune 500 list, all 500 rows.
  - `rank` Fortune 500 rank, 1-500.
  - `company` company name as printed on the list.
  - `domain` the homepage domain scored.
  - `final_url` the URL the archived fetch ended on (blank when nothing was kept).
  - `scored` 1 when the homepage was kept and scored, 0 when it was dropped.
  - `drop_reason` empty when scored; otherwise `no_record`, `no_html`,
    `status_not_200`, `short_text` or `not_the_site`.
  - `ai_mention` 1 when the visible homepage text matched the rule, blank when not scored.
  - `title_mention` 1 when the `<title>` element matched, blank when not scored.
  - `first_match_snippet` 120 characters of visible text around the first match,
    with long dashes written as ` - `.
- `groups.csv` - one row per list. `on_list` unique domains on the list,
  `scored` homepages kept, `ai_mention` count matching, `share` = ai_mention / scored,
  `title_mention` count whose `<title>` matched, `title_share` = title_mention / scored.
- `yc_by_year.csv` - Y Combinator companies by batch year, 2012-2026.
  `base` scored homepages in that year, `share` = ai_mention / base.
- `yc_by_industry.csv` - Y Combinator by the industry on the company list.
  `base` scored homepages, `share` = ai_mention / base.
- `inc5000_by_industry.csv` - Inc 5000 (2026) by industry. `share` is blank where
  `base` is under 20.
- `fortune500_by_rank_band.csv` - Fortune 500 in bands of 100 ranks.
  `scored` homepages kept in the band, `share` = ai_mention / scored.
- `inc5000_by_rank_band.csv` - Inc 5000 in bands of 1,000 ranks. Same columns.
- `METHOD.md` - the matching rule, the record selection rule, the funnel per list,
  and the limits.

Denominator everywhere is `scored`: homepages kept after the drops in METHOD.md,
never the number of companies on the list.

Per-domain rows are published for the Fortune 500 only. Inc 5000, Y Combinator
and unicorn results are published as aggregates.

Data: CC BY 4.0, Installmap, 2026-09-04.
Source: Installmap, September 2026.