More Resources
Prompt

Blend web engagement with ICP fit into one score

Tools: ChatGPT ▪ Claude ▪ Lusha

Most teams answer “who visited the website” and “who fits our ICP” with two different tools and two different tabs. This play pulls Lusha’s Website Visits data, checks each visiting company against your ICP, and folds both into a single blended score.

Typical use case: Stop working every website visit as if it means the same thing.

What is this play?

Blended Web + ICP Scoring is a Claude prompt that combines two signals teams usually track separately — how much a company engaged with your site, and whether that company actually fits your ICP — into one rating: HOT, WARM, or MONITOR.

What Lusha does

Lusha ranks companies that visited your tracked site by engagement (sessions, high-intent pageviews, recency), checks each one against your ICP and active buying signals, and returns one blended score per account instead of two separate numbers to reconcile by hand.

Best for

  • Sales — Know which visiting company is actually worth a call today.
  • Marketing — Route HOT accounts into a campaign the same week they engage.
  • RevOps — Score every source — web, inbound, event — on one consistent scale.

Turn this play into action

1

Connect Lusha to Claude

One-time setup at claude.com/connectors/lusha.

2

Set up Website Visits tracking

Add your domain at dashboard.lusha.com/account/website-visitors-settings if you haven’t already.

3

Define your ICP

Fill in the ICP fields in the prompt’s context block.

4

Run and review

Run the prompt and work the HOT list first.

The prompt

Open in Claude Open in ChatGPT 
<context>
I want to know which companies visiting our website are actually worth a rep's time — not just who visited, but who visited AND fits our ICP.

My ICP: [INDUSTRY, COMPANY SIZE, GEOGRAPHY, KEY SIGNALS]
</context>

<task>
1. Use Lusha's Website Visits data to rank companies that visited our tracked site(s) in the last 30 days, sorted by engagement score.

2. For each company in the top results, check ICP fit using Lusha's firmographic and signal data:
   - Industry, size, geography match
   - Any active buying signals (funding, hiring, exec change)

3. Combine the two into one blended score per account:
   - HOT: high engagement + strong ICP fit — work today
   - WARM: high engagement + partial fit, OR moderate engagement + strong fit — worth a look this week
   - MONITOR: low engagement or weak fit — no action yet, keep watching

4. For every HOT and WARM account, find the most likely decision-maker via Lusha and return their verified contact.

5. Return one table: Company, Engagement Score, ICP Fit, Blended Rating, Contact — sorted HOT to MONITOR.
</task>

<constraints>
- If Website Visits isn't configured for a tracked domain, say so clearly and link to setup — don't return an empty table with no explanation.
- Don't rate a company HOT on engagement alone. Fit has to be present too.
- Don't invent a decision-maker if none can be verified — flag the gap instead.
</constraints>

What you'll get back

Running this live against the demo account surfaced the right honest answer for an unconfigured domain:

“Website Visits is included in your plan, but no domain is configured yet. Once a domain is set up, you’ll be able to see which companies are visiting your site, their intent score, and session data — right here.”

That’s a deliberate part of the design — the prompt says “not set up yet,” never returns a blank table that reads like “no visitors.” Below is the shape of the output once a domain is tracked, using the exact fields Lusha’s Website Visits API returns:

CompanyEngagementICP fitRating
Nxxxhline Data88 (7 sessions, 3 high-intent pageviews)Strong — 200-person SaaS, US, funding signalHOT
Vxxakit Systems74 (4 sessions, pricing page x2)Partial — right industry, below target sizeWARM
Cxxstal Byte22 (1 session, 40 seconds)Strong fit on paperMONITOR

Illustrative — values shown once a domain is tracked; field names match Lusha’s live schema exactly.

Data used

  • Website Visits engagement score, sessions, high-intent pageviews
  • ICP fit: industry, size, geography
  • Active buying signals (funding, hiring, exec change)
  • Verified decision-maker contact

Why it works

A high engagement score and a strong ICP fit answer two different questions — “are they interested” and “should we even want them.” Scoring both together, in one number, is the only version that tells a rep what to actually do next.

FAQ

  • The prompt tells you that directly, with a link to set it up, instead of returning an empty or misleading table. A silent empty result and “not configured yet” look identical if the play doesn’t check for the difference — this one does.

  • No. The blended rating requires both. A company visiting the pricing page ten times but sitting well outside your ICP stays in MONITOR — high traffic isn’t the same as a real opportunity.

  • That skill scores inbound leads and accounts on firmographic and signal fit alone. This play adds a second input — actual site behavior — so an account that’s both visiting and a fit gets surfaced ahead of one that’s just a fit on paper with no engagement yet.

  • Up to 30 days by default, with a maximum 3-month window available. Recent activity weighs more than older activity in the ranking.

  • You’ll get the engagement ranking, but the blended score needs an ICP to compare against — without it, the prompt can only tell you who visited, not who’s worth calling.

Related Prompts

Ready to run this?

One data connection. Works in Claude, ChatGPT, your CRM, or any agent you build.

Start for free Read the Docs