Of 60 answers

The most named companies in the study, and S&P 500 financials, with how many of their 60 AI answers named them.

SWIFT47/60TRANSUNION46/60EXPERIAN42/60COINSTAR41/60NAVY FEDERAL CREDIT UNION41/60ZELLE41/60BLOOMBERG L.P.38/60TRADINGVIEW38/60BANCO POPULAR DE PUERTO RICO37/60PAYPAL37/60VENMO37/60CHEXSYSTEMS33/60EQUIFAX33/60MORNINGSTAR32/60S&P GLOBAL31/60ALPACA MARKETS30/60COINBASE30/60PINWHEEL29/60STRIPE29/60XERO29/60ACORNS28/60AFFIRM28/60PAYFAST (PAYGATE)28/60PAYONEER28/60INTERACTIVE BROKERS26/60CHARLES SCHWAB19/60THE HARTFORD9/60MOODY'S21/60CME GROUP19/60ACCENTURE11/60CHUBB10/60MORGAN STANLEY6/60AMERICAN EXPRESS14/60SYNCHRONY FINANCIAL12/60FACTSET20/60NASDAQ19/60ALLSTATE11/60WELLS FARGO18/60PAYCHEX10/60ARES MANAGEMENT3/60
Rezolve Ai
MethodDemo
Back to the study
AI visibility statementFinance AI Visibility Study 2026Issued October 14, 2026
Account holder

Castle

castle.io

Fraud, identity and complianceFinancial technology and data

Money20/20 attendee

Statement period
October 9 to 14, 2026
when the AI assistants were asked
Account no.
521 of 2,219
its rank by overall score
Place in its category
63rd of 171 fraud and identity companies
Closing balance

Named in

1of 60 answers

by 1 of 4 assistants

Statement note

What this means for Castle

AI assistants rarely name this fraud and identity company, mentioning it in 1 of 60 answers. In the same answer, they also name Sift, DataDome, and Arkose Labs. The company's overall score is 37 out of 100, ranking 521st of 2,219 in the study. Its website passes 4 out of 14 technical checks for AI tool readiness.

We asked 15 unbranded questions about its market, so no question mentioned Castle by name. AI assistants named it in 1 of the 60 answers. Counting only how often it is named, it ranks 267th of 574 in its sector. The typical financial technology and data company is named in 1.7% of answers.

Named fewer than three times in total. Read this as a sign that AI assistants rarely bring this company up, not as an exact comparison with its peers.

Overall score
37Overall score
  • Visibility2
  • Citations0
  • Sentiment100
  • Site quality87
  • AI readiness29
Summary of account
Named by AI in
1.7%
1 of 60 answers
Share of mentions
4.5%
of the mentions of it and its 5 peers
AI readiness
4/14
technical checks passed. Typical for its sector: 4
Rank in its sector
152 of 574
by overall score, where the typical financial technology and data company scores 29

Get the full report on Castle as a PDF

Enter your work email to download it.

We’ll use your email to follow up about this report. Privacy policy

Line itemsSection 01

Every question, and where the company came up

Castle came up on 1 of 15 questions. Open any window to read that answer.

  1. 01Best fraud prevention tools for protecting user authentication in US-based apps1/4
  2. 02How can behavioral analysis prevent in-app fraud for financial service platforms?0/4
  3. 03API-based fraud prevention vs SDK integrations for mobile apps0/4
  4. 04Device fingerprinting vs IP intelligence: which is more reliable for stopping fraud?0/4
  5. 05What should enterprises look for when choosing a user authentication protection platform?0/4
  6. 06How does behavioral analysis compare to bot detection for web security?0/4
  7. 07Why am I still seeing account takeovers even after deploying fraud intelligence tools?0/4
  8. 08Is investing in bot detection platforms worth it for a mid-sized SaaS business?0/4
  9. 09What is the typical pricing structure for comprehensive fraud prevention APIs in the US?0/4
  10. 10What are some non-AI approaches to preventing coordinated online abuse?0/4
  11. 11How do enterprise-grade bot detection solutions price for large-scale web applications?0/4
  12. 12What are the most effective ways to prevent account abuse for enterprise SaaS platforms?0/4
  13. 13Best practices for integrating email intelligence into user monitoring systems0/4
  14. 14Alternatives to device fingerprinting for blocking automated threats0/4
  15. 15How can false positives in automated fraud detection be minimized?0/4
Total namedOpenAI: 0/15Claude: 0/15Perplexity: 1/15Gemini: 0/15All assistants: 1/60
named the companyanswered without naming itdid not answer

Columns, left to right: OpenAI, Claude, Perplexity, Gemini. Across the study, on 50% of the questions where a company is named, only one AI assistant names it.

Transaction detailSection 02

What Perplexity actually said

An answer that names the company

Perplexitynamed the company

“Best fraud prevention tools for protecting user authentication in US-based apps”

… Castle · Web and mobile account takeover prevention · Device, browser, and behavioral signals with configurable risk rules for suspicious authentication activity.

Sift · Managed fraud operations · Account takeover, bot, and payment-fraud detection with centralized risk scoring and policy controls.

DataDome or Cloudflare · High-volume automated attacks · Bot detection and mitigation against credential stuffing, scraping, and abusive login traffic.

BioCatch · Banking and high-risk transactions · Behavioral biometrics and risk analytics suited to financial …

The AI’s answer as we received it, shortened and not checked for accuracy.

PayeesSection 03

Where the mentions went

The companies these answers named, and how the mentions were split between this company and its peers.

Also named

CompanyTimes named
  1. 01
    Sift
    5
  2. 02
    DataDome
    5
  3. 03
    Arkose Labs
    4

Share of mentions

CompanyShare
  1. Sift
    31.8%
  2. DataDome
    27.3%
  3. Arkose Labs
    22.7%
  4. PerimeterX (now HUMAN Bot Defender)
    13.6%
  5. Castle
    4.5%
  6. Human Security
    0.0%
ReferrersSection 04

Websites the AI used as sources

How many times each website was listed as a source in these answers

  1. 01
    nhimg.org
    6
  2. 02
    github.com
    6
  3. 03
    cside.com
    5
  4. 04
    radware.com
    5
  5. 05
    fingerprint.com
    4
  6. 06
    stripe.com
    4
  7. 07
    adaptivesecurity.com
    4
  8. 08
    expertinsights.com
    4
  9. 09
    datacalculus.com
    4
  10. 10
    fico.com
    3

Castle is in the study as a fraud and identity company.

By assistantSection 05

Which AI assistants named the company

Answers naming it, out of answers received

Perplexity
1/15
Claude
0/15
Gemini
0/15
OpenAI
0/15
Market comparisonSection 06

How Castle compares with all 2,219 companies

Every tower is a company in the study, taller for a higher overall score and lit where AI assistants named it. Its own group stays lit. Hover a tower to see it, and click to open it.

What makes up the score

Each part is scored out of 100.

Visibility
2

How often AI assistants name the company, and how that compares with its peers

Citations
0

How often AI uses the company's own website as a source

Sentiment
100

How positively AI describes the company

Site quality
87

How well built the website's own pages are

AI readiness
29

How easily AI tools can read and use the website

Visibility is calculated from the answers on this page. How the score is built.

Against its sectorSection 07

Compared with financial technology and data companies

574 companies in this sector. The gold tick on each bar is the typical company.

Overall score
37
Ranks 152 of 574. Typical company: 29
Named in answers
1.7%
Ranks 267 of 574. Typical company: 1.7%
AI readiness
29
Ranks 224 of 574. Typical company: 29
Against its categorySection 08

Compared with fraud and identity companies

171 companies in this category. The gold tick on each bar is the typical company.

Overall score
37
Ranks 56 of 171. Typical company: 30
Named in answers
1.7%
Ranks 85 of 171. Typical company: 1.7%
AI readiness
29
Ranks 88 of 171. Typical company: 36
AI agentsSection 09

Ready for AI agents

What this website publishes for an AI agent to read, use and pay through, on a scale of six levels.

  1. 0
  2. 1
  3. 2
  4. 3
  5. 4
  6. 5

Level 1: Basic web presence. Two of robots.txt, a sitemap and Link headers are in place.

To reach level 2:
  • Content Signals
Can an AI agent check out here?

Not by itself. Castle publishes none of the agent payment protocols we look for. An agent would have to click through the website like a person.

Agent payment protocols
  • x402Not present
  • MPPNot present
  • AP2Not present
  • ACPNot present
  • UCPNot present
Interfaces for AI agents
  • MCP Server CardNot present
  • A2A Agent CardNot present
  • Agent SkillsNot present
  • API CatalogNot present
  • llms.txtPublished
  • WebMCPCould not check
Show every check, for your technical team
Discoverability
77 of 100
  • robots.txtPassed

    A valid robots.txt with crawl rules at the site root.

    study: 86.1%
  • SitemapPassed

    An XML sitemap listing the pages, ideally referenced from robots.txt.

    study: 87.9%
  • Link headersNot met

    Send Link headers such as rel="api-catalog" or rel="describedby" on the homepage.

    study: 1.5%
  • DNS for AI DiscoveryNot present

    Advertise agent endpoints with SVCB records under _agents.

    study: 4.3%
Content accessibility
50 of 100
  • Markdown negotiationNot met

    Return a markdown version when asked with Accept: text/markdown, for example at the CDN.

    study: 6.2%
  • llms.txtPassed

    An llms.txt that tells language models what the site offers.

    study: 31.0%
Bot access control
32 of 100
  • AI bot rulesPartly met

    Add explicit robots.txt groups for the main AI crawlers.

    study: 92.7%
  • Content SignalsNot met

    Add a Content-Signal line for search, ai-input and ai-train to robots.txt.

    study: 2.3%
  • Web Bot AuthNot present

    Only relevant if you run your own agents or crawlers: publish a signing key directory.

    study: 0.3%
APIs, auth and MCP
0 of 100
  • API CatalogNot met

    List public APIs in a linkset at /.well-known/api-catalog.

    study: 1.3%
  • OAuth discoveryNot met

    Publish OpenID Connect or OAuth server metadata under /.well-known.

    study: 6.3%
  • OAuth Protected ResourceNot met

    Serve protected resource metadata at /.well-known/oauth-protected-resource.

    study: 5.3%
  • auth.mdNot met

    Publish an auth.md describing how agents sign in.

    study: 0.2%
  • MCP Server CardNot met

    Offer an MCP server and publish its server card under /.well-known.

    study: 0.8%
  • A2A Agent CardNot met

    Publish an agent card at /.well-known/agent-card.json.

    study: 0.2%
  • Agent SkillsNot met

    Publish a skills index with the main tasks agents can do.

    study: 1.0%
  • WebMCPCould not check

    Register key actions such as search or cart as WebMCP tools.

    study: 0.0%
  • ARD manifestNot met

    Publish an ai-catalog.json listing every agent interface.

    study: 0.2%

Score 36 of 100 · 97% of checks could run · checked on October 9, 2026 · castle.io · 2,200 companies scanned

See how every company did, and how this was measured.

AI readinessSection 10

The checks behind the AI readiness score

4 of 14 passed. This list is for your technical team. The percentage is how many companies in the study pass each check.

These 14 checks ran during the audit and are what the AI readiness part of the overall score counts. The scan above ran later and is stricter on Link headers and Markdown, so a check can pass here and not there.

Show the 14 checks

Discoverability

  • Passedrobots.txt publishedstudy: 85.1%
  • PassedXML sitemapstudy: 79.4%
  • Not metHTTP Link headers (RFC 8288)study: 42.5%

Content Accessibility

  • Not metMarkdown content negotiationstudy: 7.4%
  • Passedllms.txt publishedstudy: 28.1%
  • PassedToken budget (page weight)study: 97.3%

Bot Access Control

  • Not metExplicit AI bot rulesstudy: 11.5%

API / Auth / MCP

  • Not metAPI Catalog (.well-known/api-catalog)study: 1.1%
  • Not metOAuth Authorization Server discoverystudy: 5.2%
  • Not metOAuth Protected Resource discoverystudy: 4.6%
  • Not metMCP Server Cardstudy: 0.2%
  • Not metA2A Agent Cardstudy: 0.0%
  • Not metWebMCP toolsstudy: 0.3%
  • Not metAgent Skills declaredstudy: 0.1%
AI crawlersSection 11

What AI crawlers may read (robots.txt)

This website sets no rules for any AI crawler by name, like 86% of the websites whose rules we could read. Every AI crawler follows its general rules.

Training AI

  • GPTBotallowed
  • ClaudeBotallowed
  • Anthropic-AIallowed
  • Google-Extendedallowed
  • Applebot-Extendedallowed
  • Meta-ExternalAgentallowed
  • FacebookBotallowed
  • Bytespiderallowed
  • CCBotallowed
  • Diffbotallowed
  • Omgilibotallowed

AI search

  • OAI-SearchBotallowed
  • Claude-SearchBotallowed
  • PerplexityBotallowed
  • Applebotallowed
  • Amazonbotallowed
  • YouBotallowed

Answering questions

  • ChatGPT-Userallowed
  • Claude-Userallowed
  • Perplexity-Userallowed
  • DuckAssistBotallowed
  • MistralAI-Userallowed

Read from the website’s own rules file (robots.txt). An asterisk marks a crawler the file names. The rest follow its general rules. “Partly” means only some pages are blocked.

Take it with you

Get the full report on Castle as a PDF

Every number on this page in a PDF you can keep and share with your team. Enter your work email to download it.

We’ll use your email to follow up about this report. Privacy policy

End of statement
All 2,219 companiesHow we measured this
Next step

Want AI assistants to name your company?

Book a 30-minute demo to see Rezolve Ai in action and what it can do for your business.

Request a demo