It wasn’t a billboard.
It wasn’t a Meta ad.
It was Project Maxima.
HUL’s geofenced AI marketing engine, quietly mapping a 100-meter radius around every grocery store in India {HUL annual report 2025}.
Here’s the play. When a consumer’s phone enters the zone, AI pushes hyperlocal creative — calibrated to the SKU mix that store actually stocks, at the price the kirana sells it at.
The fight for the next ₹500 in your consumer’s basket is no longer in the store. It’s in the 90 seconds before they walk in.
For a Marico brand head: every top-50 outlet should be audited for geofenced opportunity by next week. For a D2C founder: what’s your 100-meter zone? Quick-commerce dark stores? Workplace cafeterias? UPI swipe locations? The category isn’t ahead of legacy media buyers — the kirana itself is.
KPI to track: CAC per SKU at outlet level. If it drops more than 15% in 90 days against control, the geofence is real. If not, you haven’t calibrated the creative or the SKU mix.
Three signals worth tracking this week
Signal 01 · FMCG — Marico is using AI for what to spend on, not what to make.
Their FY26 rollout prioritizes resource allocation — media spend, distribution intensity, SKU portfolio decisions {Marico investor call Q3}. The lens: AI for deciding, not creating. That’s the high-leverage version. Most brands are stuck on creative gen.
Signal 02 · Distribution — Nestle India now segments at the pincode level.
Not state, not city — pincode. Maggi in Indore vs Maggi in Pune are now treated as two different P&Ls {Nestle India strategy update}. The category is fragmenting faster than national brand teams can keep up. Nationwide playbooks will look obsolete in 18 months.
Signal 03 · Tech — Agentic AI was the headline at AdTech Summit 2026.
Systems that don’t just write the ad — they decide when to run it, adjust budgets in real time, pull underperforming creatives without human approval {AdTech Summit March 2026}. Nine months from now, the brand manager not using one will be visibly behind the one who is.
Operator’s Notebook
What nobody says about AI in FMCG: the bottleneck isn’t the tech, it’s the brand team’s analytical depth. Most brand managers I worked with at Marico, Nivea, Dabur spent 4 hours a week briefing agencies — and zero hours a week actually reading their own data. Agentic AI flips that ratio. The brand managers who survive the next 18 months are the ones who can read an output, not just receive a deck. If you’re a B-school student aiming for FMCG: start now. Run a Mixpanel cohort against your college canteen. The skill is portable.
P.S. — If you’re a brand or category lead trying to figure out where to start with AI in your FMCG team, reply with “AI Audit” and I’ll send the 7-question framework I use with advisory clients. Free. (The framework is also the curriculum spine for Insight Labs Academy’s AI track, launching Q3.)
—
Satyam
Operator behind The Insight Labs · Ola Electric (Head of Strategy) · ex-Marico, Nivea, Dabur, Bimbo