Benchmarks · B2B SaaS

Cybersecurity: AI readiness benchmark

46 platforms scanned on July 31, 2026. The average score is 78 out of 100. Homepages only, from a hand-picked list.

46sites scanned
78average score
55–100score range
Jul 31, 2026scan date

Key findings

  1. 1

    The average score in the Cybersecurity category is 78/100, 5 points above the all-category average of 73.

  2. 2

    Of the three pillars, Technical foundation is the strongest in the Cybersecurity category at 18/20 (90% of the maximum), and Discoverability is the weakest at 28/40 (70%).

  3. 3

    In the Cybersecurity category, Content schema (Article, FAQ…) is the most-failed check: 70% of sites did not earn full points on it.

Where the scores fall

The lowest score is 55, the highest is 100 and the average is 78, out of 100.

Fig. 1

Score range for Cybersecurity Cybersecurity sites range from 55 to 100 out of 100, averaging 78, across 46 sites. Cybersecurity · 46 platforms 78 55 100 0 25 50 75 100 AI visibility score (0–100)
The bar spans the lowest to the highest score in the category. The marker is the average.

Source: Citehound scan, July 31, 2026, 46 platforms

The three pillars

Each score is built from three pillars worth 40, 20 and 40 points. These are the category averages.

Fig. 2

Pillar averages against their maximums Average score per pillar for Cybersecurity. Discoverability, maximum 40 points: 28. Technical foundation, maximum 20 points: 18. Content & trust, maximum 40 points: 33. Discoverability · max 40 28/40 Technical foundation · max 20 18/20 Content & trust · max 40 33/40 Filled = points earned. Unfilled = gap to maximum.
Filled is the average points earned; the track is the maximum for that pillar.

Source: Citehound scan, July 31, 2026, 46 platforms

The checks sites miss most

Share of sites that did not earn full points, for the 8 most-failed checks.

Fig. 3

The 8 most-failed checks for Cybersecurity Share of sites that did not earn full points, most-failed first. Content schema (Article, FAQ…): 70%; llms.txt present: 39%; Organization / WebSite schema: 26%; Structured data (JSON-LD): 24%; Meta description: 22%; Single H1 heading: 20%; Page title: 11%; Open Graph tags: 7%. Content schema (Article, FAQ…) 70% llms.txt present 39% Organization / WebSite schema 26% Structured data (JSON-LD) 24% Meta description 22% Single H1 heading 20% Page title 11% Open Graph tags 7%
Ranked from most to least failed. A site fails a check by earning less than its full points, which includes partial credit. AI crawler access is shown by crawler in the next figure instead.

Source: Citehound scan, July 31, 2026, 46 platforms

Which AI crawlers can get in

Counts of sites, for each of 10 AI crawlers, where it is open, limited or blocked.

Fig. 4

Access by AI crawler across 46 sites Number of Cybersecurity sites where each crawler is open, limited or blocked. GPTBot: 9 open, 37 limited, 0 blocked; ChatGPT-User: 9 open, 37 limited, 0 blocked; OAI-SearchBot: 9 open, 37 limited, 0 blocked; ClaudeBot: 9 open, 36 limited, 1 blocked; anthropic-ai: 8 open, 37 limited, 1 blocked; PerplexityBot: 9 open, 37 limited, 0 blocked; Perplexity-User: 9 open, 36 limited, 1 blocked; Google-Extended: 9 open, 37 limited, 0 blocked; CCBot: 9 open, 36 limited, 1 blocked; Bytespider: 8 open, 37 limited, 1 blocked. GPTBot 9 37 0 blocked ChatGPT-User 9 37 0 blocked OAI-SearchBot 9 37 0 blocked ClaudeBot 9 36 1 blocked anthropic-ai 8 37 1 blocked PerplexityBot 9 37 0 blocked Perplexity-User 9 36 1 blocked Google-Extended 9 37 0 blocked CCBot 9 36 1 blocked Bytespider 8 37 1 blocked Open Limited Blocked
Open: no rule keeps the crawler out. Limited: at least one Disallow rule applies to the crawler, mostly an ordinary path such as /wp-admin/, and the rest of the site is open to it. Limited is not a block, although the scan gives it half credit. Blocked: the whole site is disallowed.

Source: Citehound scan, July 31, 2026, 46 platforms

Method and limits

What we scanned

46 platforms in this category completed a scan on July 31, 2026. 4 more on the list could not be reached or timed out. They are excluded from every figure on this page.

What we pulled

For each site: robots.txt, sitemap.xml, llms.txt and the homepage, retrieved live. Homepages only. Inner pages were not scanned, and each site was scanned once.

What we scored

One score out of 100 per site, split across three pillars: discoverability (40), technical foundation (20), content & trust (40). The checks and their point values are on the methodology page.

How the sites were chosen

The lists are hand-picked, well-known sites in the category. They are not a random sample, so these figures describe the sites we scanned, not the category as a whole.

Repeat scans

In one October 2026 rescan of the same lists, category averages moved by one point or less. Individual sites moved more: about a third changed by 3 to 15 points.

What this measures

Whether AI systems can reach a site, and whether its homepage carries the signals machines read. It does not measure whether any assistant names a brand in an answer. Model responses vary by prompt, session and training data, and no scan of public files can predict them.

What this doesn’t tell you

  • Anything about pages other than the homepage.
  • Why the unreachable sites could not be scanned, or how they would have scored.
  • Whether any of these signals relates to being named in an AI answer. This page measures inputs, not outcomes.

How to cite this

Citehound, “AI readiness benchmark: Cybersecurity,” July 31, 2026. https://getcitehound.com/benchmarks/cybersecurity

Canonical URL: https://getcitehound.com/benchmarks/cybersecurity

To see where your own site lands against these numbers, run the free scan. All benchmarks.

Scan your site free