Case Study 05 · CRM & AI Sales Infrastructure
A custom CRM and AI lead scoring system that saved ₹15.1 lakhs in annual tech costs
The brief
Cafemutual's sponsorship sales ran on relationships. AMCs, event sponsors, advertising partners. Complex deals, long cycles, multiple stakeholders. The kind of sales process that absolutely needs a structured CRM, and almost never has one.
I built the entire system from scratch. Tailored to how their team actually worked. At ₹9,050/month instead of ₹15,000+ for HubSpot.
At a glance
Cafemutual Sponsorship Sales
B2B sponsorship deals with AMCs, event sponsors, and advertising partners. Long cycles, multiple stakeholders, relationship-driven. Starting state: deals lived in spreadsheets, inboxes, and memory.
The challenge
No CRM. No pipeline visibility. No lead scoring. No automated follow-up. Deals going quiet by accident. Revenue stalling for no visible reason.
No pipeline structure
Deals lived in spreadsheets, inboxes, and memory. Leadership had no way to see where revenue was coming from or where deals were stalling.
No lead prioritisation
Without lead scoring, the sales team treated every lead equally. Hot leads got the same attention as cold ones. Missed follow-ups were costing real deals.
No outreach automation
Personalised outreach was written from scratch for every lead. In a relationship-driven B2B business, a poorly timed or generic email can cost a deal.
The solution
Designed an 18-field CRM built around how the team actually works.
Contact info, deal details, timeline tracking, relationship history. Pipeline across four stages: Proposal, Negotiation, Closed Won, and Lost. Built to reflect reality, not a vendor's ideal workflow.
Built an AI lead scoring model with 6 weighted signals.
Past event history (30%), deal size (25%), engagement frequency (20%), sector relevance (15%), recency of contact (10%). Leads auto-classified as Hot, Warm, or Cold. The team always knew who to call first.
Automated follow-up reminders and personalised email drafts.
No deal goes quiet by accident. AI-generated personalised sponsor email drafts so outreach was fast, relevant, and consistent, without sounding templated.
Designed full-stack sales infrastructure blueprint.
Integration roadmap: Fireflies AI for call notes, Apollo for lead enrichment, Clay for web intelligence. Executive dashboard covering pipeline overview, lead owner performance, and category analysis.
Results
Key takeaways
01
A CRM is only as good as the process it reflects. Build it around how your team actually works, not how a vendor thinks you should, and adoption is immediate.
02
Lead scoring removes guesswork. Your best salespeople already do this intuitively. Building it into the system means everyone operates at that level.
03
Build vs. buy is a real strategic decision. Sometimes the right answer is neither a full platform nor a spreadsheet, but something custom-built for your exact workflow.
Sales system
needs fixing?
A well-built CRM changes how your whole team operates, not just where you store contacts.
Connect on LinkedIn ↗