Skip to main content
When AI agents like GPT, Claude, or Perplexity research a product on behalf of a buyer, they leave a trail — crawling your site, running evaluate queries, and in some cases booking demos directly. ClientX Visitor Intel captures all of that activity, groups it by company, scores each company’s intent level, and surfaces them in a feed your team can act on immediately.

What Visitor Intel Shows

Each row in the Visitor Intel feed represents a company whose AI agents have interacted with your product. For every company you can see:

Company & Domain

The company name and domain resolved from the AI agent’s activity. Companies are grouped so all agent interactions from the same domain appear together.

Intent Level

A three-tier classification — Demo Booked, High Intent, or Research — based on the type and volume of AI agent activity observed.

AI Interactions

A breakdown of activity type: crawl events (the agent fetched your content), evaluate queries (the agent ran a structured product fit assessment), and demo bookings (the agent completed a booking on behalf of the buyer).

ICP Fit Score

A percentage score reflecting how closely the company matches your Ideal Customer Profile, calculated from the evaluate query content and company firmographics.

Use Case

The specific use case or problem the AI agent surfaced during its evaluation — extracted from the evaluate query so you know exactly what the buyer is trying to solve.

Last Seen

How recently the company’s AI agent was active. Rows are sorted by recency so the freshest buying signals always appear at the top.

Intent Levels

ClientX classifies every company into one of three intent levels based on what their AI agent actually did:
LevelWhat triggered itWhat it means
Demo BookedThe AI agent completed a demo bookingThe buyer instructed their AI to schedule a meeting — the highest possible signal
High IntentThe AI agent ran one or more evaluate queriesThe buyer asked their AI to assess product fit — they are actively evaluating
ResearchThe AI agent crawled your contentEarly-stage discovery — the buyer is gathering information
Evaluate queries are structured requests where an AI agent assesses whether your product fits a specific use case. They are a strong buying signal because they require the buyer to have explicitly asked their AI to evaluate your product.

Summary Statistics

At the top of the Visitor Intel page, four KPI cards give you an instant snapshot of AI-driven buying activity for your selected time range:

Companies

The total number of distinct companies whose AI agents have interacted with your product in the selected period.

AI Interactions

The combined count of all crawl events, evaluate queries, and demo bookings across all companies.

High Intent

The number of companies that ran at least one evaluate query — your highest-priority outreach targets.

Demos Booked

Calendar bookings completed by AI agents on behalf of buyers during the period.

Filtering the Feed

Use the filter tabs above the feed to narrow the list by intent level:
  • All — every company with any AI agent activity
  • Demo Booked — companies where an AI agent completed a booking
  • High Intent — companies where an AI agent ran an evaluate query
  • Research — companies in early-stage discovery mode
The time range selector (7 days, 30 days, 90 days) controls how far back the feed looks.

Turning Companies into Outreach

When you spot a high-intent company in the feed, you can launch a personalised outreach sequence directly from Visitor Intel without leaving the page.
1

Find a high-intent company

Filter the feed by High Intent or Demo Booked to focus on companies with the strongest buying signals. Review the use case and ICP fit score to confirm relevance.
2

Click 'Reach Out'

On any company row, click Reach out. This opens the outreach drawer, which pre-fills the company name, domain, intent level, ICP fit score, and use case context extracted from the AI agent’s evaluation.
3

Add a contact email

If a contact email was captured from the session it will be pre-filled. If not, click Find Decision Makers to search your connected Apollo.io or Hunter.io account for contacts at the company. Select the right person from the results.
4

Start the sequence

Click Start Outreach Sequence. The AI Outreach Agent writes a personalised cold email sequence referencing the company’s research activity and ICP fit, then queues it for sending.
Find Decision Makers requires a connected Apollo.io (Basic plan or higher) or Hunter.io account. If neither is connected, go to Settings → Integrations to add one before attempting outreach from Visitor Intel.

Agentic Endpoints

For AI agents to discover and evaluate your product, ClientX publishes a set of agentic endpoints on your domain. These are created automatically when you set up your workspace — no configuration is needed on your end.
Your agentic endpoints (/llms.txt, /evaluate, /capabilities.json) are live as soon as your workspace domain is configured. When AI agents crawl or evaluate your product through these endpoints, the activity appears in your Visitor Intel feed automatically.
To verify your endpoints are discoverable, check Settings → Workspace and confirm your agent domain is set. Once it is, any AI agent that evaluates your product will appear in this feed.