AI Lead Scoring Engine — TechVista Software
Client
TechVista Software, a Pune-based software product company selling B2B SaaS to mid-market enterprises.
Problem
The sales team received hundreds of inbound leads a month — demo requests, trial sign-ups, and website enquiries — and had no objective way to know which were worth chasing. Sales reps followed up in the order leads arrived, wasting hours on tyre-kickers while genuine enterprise buyers went cold. Management knew the data to prioritise leads existed in the CRM, but nobody could turn it into a ranked daily list.
Solution
Gravitas India built a machine-learning lead scoring engine on top of TechVista's existing CRM. The model learned from seven years of historical lead-to-close data — firmographics, engagement behaviour, source, and sales cycle length — and produced a 0–100 purchase-likelihood score for every inbound lead in near real time. Sales reps received a daily ranked queue instead of an inbox pile.
Implementation / Technical Approach
- Feature engineering — 40+ signals distilled from CRM history: industry, company size, page visits, demo attendance, email engagement, source channel, and time-to-first-touch.
- Model selection — gradient-boosted classifier trained on labelled historical outcomes, with precision tuned at the top of the funnel where rep time is spent.
- Explainable scores — every lead's score carried the top contributing factors, so reps understood why a lead ranked high instead of trusting a black box.
- CRM-native delivery — the daily ranked queue rendered inside the existing CRM as a scheduled report, with no change to the sales team's workflow.
- Feedback loop — closed-lost and closed-won outcomes re-fed the model monthly, so scoring improved as the sales cycle evolved.
Technology
Machine Learning Gradient Boosting CRM Integration Automated Reporting
Outcomes
With reps working the ranked queue, the close rate on scored leads rose 29%, follow-up time dropped 41%, and revenue per rep more than doubled within two quarters. The scoring engine also surfaced a pattern no one had noticed: enterprise leads from partner referrals converted at 3× the rate of cold inbound — which reshaped TechVista's channel strategy entirely.