How XPndAI built a multi-channel AI lead agent — website chat, WhatsApp, and email — that qualifies, scores, and routes leads 24/7 for a Delhi NCR B2B SaaS company.
The company had a solid product and growing inbound interest — roughly 180–220 website leads per month from organic search, LinkedIn campaigns, and Google Ads. The problem was what happened after a lead submitted the form.
Their two-person sales team was manually reviewing every lead the next morning, calling or emailing to ask qualification questions, then deciding whether to proceed. By the time a lead got a substantive response, 18–36 hours had passed. For B2B SaaS buyers who are evaluating 3–5 vendors simultaneously, this delay was costing them deals they never knew they lost.
"We were spending 60% of our sales day on discovery calls with leads who had no budget or were too small for our platform. We needed the first filter to happen before a human ever got on a call." — VP Sales
XPndAI designed and built a three-layer AI lead qualification system that operates across every channel where leads arrive.
A context-aware chat widget on the website that engages visitors who show high-intent signals (pricing page, features page, 60+ seconds on demo page). The agent introduces itself, asks a structured qualification sequence (company size, current tool, team structure, timeline, budget range), and produces a lead score in real time.
Leads who submit the contact form receive an immediate WhatsApp message from the company's business number. The agent continues the qualification conversation over WhatsApp — the medium B2B buyers in India actually respond to — collecting remaining BANT data if the website chat didn't complete it.
For leads who don't engage on chat or WhatsApp within 4 hours, the email agent sends a personalized follow-up with a one-click qualification survey. Responses are parsed and scored automatically.
GPT-4o with a custom system prompt encoding the company's ICP, qualification rules, and product knowledge base.
Custom scoring model (0–100) based on company size, budget, timeline, decision-maker authority, and CRM signals.
Bi-directional HubSpot integration — leads auto-created, enriched with qualification data, scored, and routed to the right sales rep.
Meta WhatsApp Business API via Twilio — real-time conversation with automatic handover to human when needed.
Triggered email flows for non-responders with personalized context pulled from earlier website visit data.
When a lead hits the qualified threshold, sales rep receives a Slack message with a full briefing: company, BANT status, pain points mentioned, and conversation transcript.
Reviewed 6 months of closed-won and closed-lost CRM data to identify what separated qualified from unqualified leads. Built the qualification ruleset and scoring weights with the VP Sales.
Website chat agent built and deployed on staging. Sales team ran 30 test conversations to calibrate language, qualification question order, and fallback handling.
WhatsApp Business API integration live. HubSpot sync built — leads created, scored, and deal stage updated automatically. Slack notification flow configured.
Email follow-up sequences built. Full multi-channel flow tested end-to-end with 50 synthetic leads across all channels. Edge case handling refined.
Gradual go-live: 30% of leads routed through agent first week. Monitored conversation quality daily. Full 100% traffic by Day 10.
"The AI handles the first 80% of the sales conversation now. By the time our reps get on a call, they already know the company size, the pain, the budget, and who else is involved in the decision. Close rates went up because we're doing better discovery — we're just not the ones doing it." — VP Sales, 90-day review
Book a 30-minute call. We'll map out your qualification rules, channels, and CRM integration — and give you a fixed-price estimate before we write a line of code.