AI Lead Scoring
A list of names is not a pipeline. A ranked list is.
Describe the companies and people you want in one paragraph. IDview scores every identified visitor 0–100 against it — and tells you why.
What goes into the score
Your ICP, in plain English
No rule builders. Write “Series B SaaS in North America, 50–500 people, RevOps or demand gen leaders” and the model takes it from there.
Company intelligence
Industry, headcount, revenue band, tech stack and a live crawl-based intel report from Bytemine — cached, so you never pay twice for the same account.
Behavior that matters
Pricing views, docs depth, repeat sessions, time on page and return velocity. A second visit this week outranks a first visit last month.
Person-level fit
Title, seniority and function are weighed against who actually signs off in your deals — not a generic B2B template.
Three buckets, zero debate
Hot · 80–100
Strong ICP match with real buying behavior. Call today, not next sprint.
Warm · 50–79
Fits the profile but signal is thin. Worth a sequence, not a cold call.
Cool · 0–49
Wrong size, wrong market or wrong role. Filtered out so your reps stay focused.
Example ICP
“B2B SaaS companies in the US and Canada, 50–500 employees, selling to marketing or revenue teams. Buyers are Directors and VPs of Demand Generation, RevOps or Growth. Skip agencies, students and job seekers.”
Built to be used, not configured
Generate the ICP with AI
Do not have one written down? Point IDview at your own domain and it drafts a profile from your positioning, then you edit it.
One ICP per tracker
Different sites, different buyers. Every tracker carries its own profile and scores independently.
Reasoning you can audit
Each score ships with a short written rationale, so your team trusts the number instead of arguing about it.
Sorted feed, ranked exports
The visitor feed sorts by score by default, and every CSV export carries the score and reasoning with it.
Straight answers
How is this different from a points-based scoring rule?
Point rules break the moment reality changes shape. The model reads your ICP as language and compares it to enriched firmographics and real behavior, so a 40-person agency and a 40-person fintech do not score the same just because they share a headcount bucket.
Can I change my ICP later?
Any time. Update the description and new sessions score against it immediately.
Does it score anonymous visitors?
It scores anything we can resolve to a company. Fully anonymous sessions stay visible in the feed but are not ranked — we will not invent a score from nothing.
Do agent conversations affect the score?
Yes. If the AI Website Agent talked to the visitor, intent from that conversation feeds straight into the score.
Rank, then act
Stop working your traffic in the order it arrived.
Write one paragraph. Get a ranked buyer list on your first session.