Case Study 03 · AI & Marketing Automation
AI marketing systems that cut CPC by 59% and saved ₹5 lakhs a year
The brief
Most companies talk about AI transformation. Very few actually build it, measure it, and make it stick inside the organisation.
During my time at Cafemutual, a B2B media platform serving India's mutual fund distributor ecosystem, I built three AI systems from scratch. Working systems. The kind that saved 13–16 hours of manual work every single day and cut technology costs by over ₹5 lakhs a year.
This is the marketing workflow automation piece.
At a glance
Cafemutual
India's leading B2B financial media platform. 3 lakh+ MFDs, RIAs, wealth managers, and AMCs. Events: SIF Summit, Beyond Borders, Confluence Passives Conference. Problem: entirely manual marketing operation.
The challenge
10+ Meta campaigns, manual lead outreach, no clean database, no analytics feedback loop. Every decision was made on instinct. The operation couldn't scale.
No segmentation or retargeting
Ad campaigns ran without clear audience segmentation. CPC was high, conversions were low, and there was no feedback loop showing what was working.
Manual lead outreach
Reaching 500+ MFD leads for events like SIF Summit required days of manual phone calls and email writing. The entire team was tied up before the campaign even started.
No clean database
Incomplete, unverified contact records. Outreach was limited by data quality. Without a working analytics setup, there was no way to improve.
The solution
Built Meta ad infrastructure from scratch.
10 Meta ad campaigns across SIF Summit 2026, Beyond Borders 2026, and Confluence Passives Conference. Set up Meta Pixel on both domains, detailed audience segmentation, retargeting audiences, and lookalike targeting off high-intent visitor behaviour.
Automated email outreach with Python + Google Apps Script.
350 personalised emails in a single run. What used to take 6 hours of manual work per day happened in minutes. Scaled outreach without scaling headcount.
Built a verified database of 1,500+ SIF distributors.
Python web scraping + ARN matching against official AMFI data. Clean, verified, ready for outreach. No more guessing who to contact.
Personally ran cold calling at 20% RSVP conversion.
Cold-called 550+ MFD leads, maintained a 20% RSVP conversion rate, and directly contributed to 98 SIF Summit registrations. At ₹7,000 per ticket, that's ₹6.86L in direct revenue.
N8N automation workflows for reporting and pipelines.
Scheduling, reporting, and data pipelines automated via N8N, saving an additional 4 hours of daily manual work across the marketing team.
Results
Key takeaways
01
Automation doesn't replace strategy, it amplifies it. If your targeting is wrong, automating it at scale just wastes money faster.
02
Building your own tools at a fraction of the cost of commercial platforms is possible if you're willing to put in the technical work upfront.
03
A clean database is the foundation of everything. Bad data makes every campaign, every outreach, and every decision worse.
Want this built
for your team?
Marketing automation that actually works takes system design, not just tools.
Connect on LinkedIn ↗