Every lead, a number. Every number, a reason.
A 100-point predictive scoring engine that evaluates every lead across five dimensions — and shows you exactly which signals added or removed points. Built on 69M real CRM events.
Most lead scoring is a black box. Ours isn't.
Other tools give you a number with no explanation. You can't coach with it, can't tune it, can't trust it. Here's what's actually broken across the industry.
Built differently — and shipping today
Five dimensions, itemized, tunable, and grounded in 69 million CRM events.
Five-dimension score, every signal itemized
Each lead's score breaks down across Speed (20pt), Engagement (25pt), Intent (30pt), Freshness (15pt), and Agent Activity (10pt). You see every signal that contributed — tour requests, valuation inquiries, email clicks, response time — with its point value attached.
Real-time recalculation as signals arrive
When a lead opens an email, saves a listing, or replies to a text, the score updates in real time. No nightly batch jobs, no stale numbers. Your priority queue is always current.
Industry-tuned weights — no generic defaults
Real estate teams get tour-intent and valuation-inquiry weights tuned from 69M events. Agencies get retainer-cycle signals. SaaS teams get response-time-to-MQL. The model adapts to your vertical.
Tunable, with a visual Scoring Lab
Admins can adjust signal weights, set per-category thresholds, and run what-if simulations against historical pipeline data — all without touching code. Ship a tuned model in an afternoon, not a quarter.
Why this model is structurally different
Who uses Pocket Isa scoring
Admin / Owner
Tune the model, ship a version your team trusts.
- Visual Scoring Lab — no code
- What-if simulation against historical pipeline
- Per-category thresholds and weights
Manager
Coach with specifics. The score is no longer a black box.
- See which signals lifted (or hurt) each lead
- Spot agents whose leads consistently underscore
- Compare cohort scoring across windows
Agent
Walk in, see the queue, work top-down.
- Score visible on every lead
- Click any score to see why
- Sorted by close probability × deal value
Pulls signals from the tools you already use
Scoring works on data ingested via Zapier, our REST API, or CSV — no rip-and-replace required.
Built on 69.1 million real CRM events.
Most scoring models are trained on synthetic data or one CRM's narrow signal set. Ours is trained on 4 years of behavior across 5 platforms.
Read the methodologyEverything teams ask before signing up
What signals does the score use?
Speed (response time, time-to-first-touch), Engagement (email opens/clicks, listing views, saved searches), Intent (tour requests, valuation inquiries, pricing questions), Freshness (recency of activity), and Agent Activity (calls logged, replies, follow-up completion).
Can we change the weights?
Yes. Admins can adjust weights per category in the Scoring Lab. Changes can be simulated against historical pipeline before shipping live.
Why is each signal point value itemized?
Because opaque scores get ignored. When a manager can show an agent why this lead scored 72 (not 65), the agent trusts the queue. That's the difference between a score that drives behavior and one that doesn't.
How fresh is the score?
Real-time. When a lead acts, the score updates. No nightly batch.
Can we score leads from any CRM?
Yes. Pocket Isa ingests events via Zapier, REST API, or CSV upload from any CRM. The auto-mapping engine handles 100+ header variations from the major platforms.
What if a signal doesn't apply to my industry?
Industry templates ship with the relevant dimensions pre-weighted. You can disable signals that don't matter for your vertical and the score will renormalize to 100.
How accurate is it compared to other models?
Accuracy depends heavily on training data. Our edge is the 4-year cross-CRM dataset — competitors trained on one platform's data fundamentally underperform on cross-platform behavioral patterns.
Score every lead in your pipeline this week
Connect your CRM via Zapier or CSV. We'll have your first scored queue inside 48 hours.