RGM·AI — the objection was the architecture
An agentic pricing and revenue-growth-management product for CPG. Three enterprise buyer objections were stalling deals; one design choice removed all three.
- Role
- Product Manager & GTM Lead
- Org
- TCS — AI SaaS & Enterprise Solutions
- Timeline
- 2024 – 2025
- Result
- 1st enterprise deployment + 2 follow-on deals
Three objections, repeated in every room
The problem
Buyers evaluating the product raised the same three things: data residency, vendor lock-in, and "we already invested in a model." Deals stalled there.
They didn't want another dashboard
Discovery
JTBD interviews with Revenue Management Directors and Commercial Finance leads at eight target accounts, plus a competitive analysis of twelve incumbents. The finding: buyers wanted decision support, not more reporting.
- Reporting → recommendation engine.
- Generic analytics → CPG-vertical models.
- Desktop-first → field-mobile.
Compose on their model, not ours
Architecture
Instead of shipping a proprietary model, the product composed on the customer's own LLM via RAG over their own data. Residency, lock-in, and sunk model investment all stopped being objections in a single design choice.
The PMF breakthrough was realising our buyers didn't want more data — they wanted fewer decisions. Every product choice flowed from that.
Monitor and recommend, never act
Governance
- No autonomous action — a human approval gate sits on every consequential recommendation.
- Explainability was built to serve that gate: interrogate a recommendation in seconds rather than minutes.
- Success metrics: recommendation acceptance rate, decision-time reduction, adoption tracked separately by role (rep / manager / trade-promo manager), and margin and revenue attributable to accepted recommendations.
Selling it
Go to market
ICP: $1B+ CPG companies running manual trade spend, with the Revenue Management Director as the budget holder. I built the battlecards, ROI calculators, and demo scripts, and led C-suite discovery and executive presentations.
Result
- First enterprise deployment.
- Two follow-on deals.
- 2024 Stevie Award — AI Product of the Year.