A lead's score lives in real time. Every signal captured, every classified reply, every agent pass over the base, every adjustment to your ICP, and the engine recomputes. You get two separate scores · alignment to your target (how well the prospect fits your ICP) and actual warmth (how warm they really are). On every lead, you see the full breakdown with the % match per criterion. No black box.
The alignment score measures relevance to your ICP (industries, titles, sizes, geo, stack, keywords). The warmth score measures actual engagement (market signals, replies, opens, clicks, detected intent). A highly-aligned but silent lead isn't hot. A very-engaged but less-aligned lead can still be worth the effort. You set the threshold on each.
ICP alignment, market signals (15+ types · funding rounds, hires, role changes, tech-stack additions, LinkedIn engagement, form submissions, etc.), reply intents (call request, closing request, positive, question, objection, negative, unsubscribe), the boost the AI gives when a message really landed, and a stage floor that keeps booked leads at the top.
The score recomputes on every event that touches the lead · a new signal landing, a reply coming in, an agent revisiting the base, a change to your ICP. Latency between the event and the updated score is measured in seconds. No batch, no cron to monitor.
On every prospect's detail page, you see exactly which of your ICP criteria they match (with the fit percentage per criterion) and which they miss. You also see the warmth detail · how many signals counted, which replies pushed the score up, which intents were detected. You always know why the score is what it is.
Old signals lose weight automatically, on a 30-day decay. Past 90 days, a signal has zero influence anymore. No stale signal polluting a hot lead's score today, no cleanup job to schedule.
You can run several ICPs side by side (by segment, by product, by campaign). Every lead is tested against all your active ICPs and gets the best-match score. Edit any ICP → the relevant base auto-rescores, no silent drift.
No. The ICP gives the alignment score (relevance to your target). The warmth score, on the other hand, fuses market signals, AI-classified reply intents, the boost the AI grants for messages that really landed, and a stage floor (a booked lead is necessarily warm, for example). Both axes are visible separately on every lead.
On every event that touches it · a new signal landing, a reply classified by the AI, an agent pass over the base, a change to your ICP. No nightly batch, no cron to monitor.
Each agent has a minimum threshold (70 by default). Any prospect whose warmth crosses that threshold is automatically enrolled in the agent's sequence or adaptive plan. You can tune the threshold per agent. A 'hot lead' notification also fires whenever a lead crosses the qualification threshold you set for the workspace.
The engine kicks off auto-rescore across the relevant base · every lead is re-evaluated against all your active ICPs. No silent drift, no risk of a lead staying mis-scored after a targeting change.
Yes. A classified negative reply lowers the warmth (an unsubscribe drops it sharply). And with time decay, a lead very hot 60 days ago but silent since sees its warmth fade naturally, until a new signal or reply reignites it.
Trial on request · contact our sales team.