Monte Carlo is building its inbound playbook from scratch and sells trust in AI agents. But when we fetched montecarlo.ai on November 4 with a plain HTTP client and again with an AI assistant's web-fetch tool, every page (homepage, /agent-trust, the blog, the pricing request page, the sitemap, even /llms.txt) came back as the same "One moment, please..." verification page with no content. A buyer who asks an AI assistant "how does Monte Carlo Agent Trust work?" gets an answer built from other sites, and a site scanner reads the homepage as having no H1, description or share image. Real browsers see the page fine (H1 "Trust your agents in production").
First action
Fetch /agent-trust with curl and with one AI assistant; if both get "One moment, please...", open the bot-management rule.
How to measure it
Key pages readable by AI fetchers
- Now
- –
- Target
- 100% of top 50
- By
- 30 days
Steps
Reproduce
Fetch ten key pages with curl and two AI assistants; record what each gets.
Web team 1 hour
Find the rule
Locate the bot-verification rule in front of the site and its allow list.
Web team 1 day
Allow good readers
Let verified search, AI-assistant and link-preview agents through; keep the challenge for abusive traffic.
Web team 2 days
Open machine files
robots.txt, sitemaps and a new llms.txt always load without a challenge.
Web team 1 day
Write llms.txt
Agent Trust, data observability, pricing, customers and integrations, in that order.
Product Marketing 1 day
Test answers
Ask three assistants what Agent Trust does and compare with the page.
Product Marketing 1 day
Check previews
Paste key URLs into Slack, LinkedIn and email to confirm title and image show.
Marketing Operations 1 hour
Monitor
Weekly check that key pages return content to plain clients.
Web team weekly
Why it works
Change the bot rule so that known search crawlers, AI assistant fetchers and link-preview bots get the page, while the challenge stays for abusive traffic. Exempt robots.txt, llms.txt, the sitemaps and the blog first. Publish an llms.txt on montecarlo.ai like the one the docs already have, with Agent Trust, pricing and customer pages at the top. Then test with the same tools buyers use: ask three AI assistants to read the Agent Trust page and check what they quote.
Enterprise data and AI leaders now research vendors through AI assistants before they talk to sales. If those assistants cannot read the site, the category Monte Carlo is defining gets described by Arize, Braintrust and Bigeye instead, and the new SDRs inherit fewer, colder inbound leads.
Ready to use
ChecklistMachine-readable site check
[ ] curl https://montecarlo.ai/ returns the real H1 [ ] /agent-trust, /blog/ and /request-for-pricing/ return content [ ] /sitemap_index.xml returns XML [ ] /llms.txt exists and lists Agent Trust first [ ] Link previews show title, description and image [ ] Two AI assistants quote the Agent Trust page correctly [ ] Demo form still protected
Risks and how to handle them
Scrapers get through too
Allow by verified agent and rate limit, not by opening everything.
Rule changes break forms
Keep the challenge on form posts and test the demo form after the change.
Evidence
- Your websiteChecked November 4, 2026 with a standard HTTP client and again with an AI assistant's web-fetch tool: the homepage, /agent-trust, /blog/, /request-for-pricing/, /sitemap_index.xml and even /llms.txt all returned the same "One moment, please... Please wait while your request is being verified" page (HTTP 200, about 12 KB, no content). Only robots.txt loaded.Source
- Site scanprogrammable.info's scan of montecarlo.ai (November 4, 2026) also received only the verification page, so it recorded the homepage headline as "One moment, please..." with no meta description, no H1 and no social share image; a year earlier the same scan read the headline "Data + AI Trust You Can Scale".Source
- HiringStrategic Sales Development Representative (remote Americas, posted August 25, 2026): "triage inbound leads against our best-fit customer profile, outbound to our most strategic prospects, follow up fast"; "This isn't just a lead-routing role - you'll help build the inbound playbook from scratch, working hand in hand with marketing to hit our pipeline goals." The SDR will "use AI tools daily to research accounts, personalize outreach at scale, prioritize your queue".Source
- Google PageSpeedIn a real browser the homepage loads fine: H1 "Trust your agents in production", mobile score 96/100 (desktop 100), 524 KB over 58 requests, about 1 second of render-blocking requests, built on WordPress with WP Rocket and a Qualified chat widget. Real Chrome visitors pass Core Web Vitals (largest paint 2.2 seconds, interaction 152 ms, layout shift 0.01).Source
- Your docsdocs.getmontecarlo.com loads for every reader and publishes an llms.txt listing 492 pages: Agent Trust guides (Google ADK, Snowflake Cortex Agents, Databricks Agent Bricks and AI/BI Genie, Reinforcement Loop, conversation clusters, agent evaluation monitors), an MCP server, and onboarding examples done "with Claude". Its one-line description still reads "Data breaks. We ensure your team is the first to know and the first to solve."Source
- Competitor researchRivals' own homepages: Bigeye, "Data Observability and AI Trust Platform"; Arize, "Agent Observability, Evaluation & Improvement Platform"; Braintrust, "The active observability platform for agents"; Acceldata, "Autonomous Data & AI Platform for Hybrid Clouds"; Anomalo in data quality.Source