Agentic AI in revenue growth management: From hype to decision intelligence
This post is co-authored by Asper.AI, Chief Product and AI Officer, Soudip Roy Chowdhury, and RGM Business Unit Lead, Vibhor Mishra
Revenue growth management (RGM) has never been more essential—or more difficult to execute well.
For years, many consumer goods companies could rely on a relatively stable set of playbooks: predictable shopper behavior, consistent channel economics, and promotional mechanics that reliably delivered results in a more stable environment. Consumers are increasingly price-aware and deal-oriented. Digital platforms make comparison shopping effortless, and agentic commerce accelerates the journey from intent to purchase, while the margin-volume equation continues to shift. In short: what used to be good enough in pricing, promotions, assortment, and trade investment is now a structural risk.¹
At the same time, the broader fast-moving consumer goods (FMCG) model is under pressure. Industry incumbents are navigating slower demand, a reshaping of channels, erosion of traditional scale advantages, and the relentless rise of digitally enabled business models.² The stakes are clear: RGM is no longer a specialized capability sitting inside sales or finance. It is becoming the connective tissue between growth strategy and execution.
However, many RGM organizations continue to operate with fragmented systems, inconsistent definitions, and analytics that struggle to keep pace with change. In a recent discussion I had with leaders from Asper.AI, Chief Product Officer Soudip Roy Chowdhury and RGM Business Unit Lead Vibhor Mishra, we went straight at this reality—and what it will actually take for Agentic AI to deliver outcomes in RGM, rather than headlines.
Why RGM has to change (and why the timing is urgent)
Boston Consulting Group (BCG) recently argued that, amid economic uncertainty, consumer companies must shift their RGM bias from higher profits and productivity to higher volume and market share—and that winners will master three challenges: winning shopper missions, cross-functional orchestration, and rebuilding infrastructure on AI-enabled tools.³
That framing resonates because it reflects what I see in the field: shoppers are changing faster than conventional processes can interpret, and traditional analytic cycles are too slow for today’s volatility. The opportunity is real—but only if we confront the operational reality inside many organizations:
- Trade and spend decisions managed through disconnected tools (sometimes Excel and email).
- Siloed dashboards built by business-unit fringes because there is no shared platform.
- Inconsistent key performance indicator (KPI) definitions across teams, markets, and retail customers.
- A shortage of scalable decision support for complex trade-offs (price versus volume, promo ROI versus. brand equity, distribution versus mix).
If we don’t fix these foundational issues, agentic narratives risk turning into overly optimistic technology narratives—where the technology story races ahead of the business systems required to benefit from it.
Agentic AI, demystified: It’s not just better search
One of the most helpful parts of my conversation with Soudip Roy Chowdhury was his crisp distinction between vanilla retrieval and what truly makes an agentic system useful in RGM.
As he described it, the differentiator is grounding beyond data—combining domain knowledge with organizational knowledge and role-based interpretation. That means capturing not only what the metric is, but how different people in the organization use it to make decisions.
Asper.AI grounds RGM insights on domain, organizational, and role knowledge. Therefore the insights retrieved for a CFO is different from a Head of RGM, because their KPIs are very different. This approach increases the utility of an agentic system than just vanilla retrieval systems.
Soudip Roy Chowdhury, Chief Product Officer, Asper.AI
This matters enormously in RGM because success is not about producing “an answer.” It’s about navigating trade-offs across levers—pricing, promotions, assortment, trade terms—while reconciling the differing objectives of sales, marketing, finance, and category teams.
In the discussion, Soudip Roy Chowdhury explained how role-specific grounding can live in a knowledge base (for example, a domain ontology in the form of a graph for knowledge organization and reasoning) that maps KPI meaning, data sources, and how business entities relate—enabling agents to respond with nuance rather than generic output.
The RGM foundation: From System of Record to System of Intelligence to Agentic AI
Then came a moment I loved—because it turned a complex topic into an executive-ready mental model.
Vibhor Mishra described the prerequisites for an RGM assistant as two foundational layers:
- System of Record: The authoritative source of spend decisions, financial data, and account-level profit and loss truth.
- System of Intelligence: The ability to bring data together, standardize mappings/assumptions, and operationalize analytics and models (elasticity, forecasting, simulation).
This is the reality check: Agentic AI cannot compensate for missing financial truth, fragmented trade data, or absent governance. It can accelerate and augment—but it cannot conjure decision-quality inputs out of thin air.
At the same time, Vibhor Mishra offered an important nuance: organizations don’t have to finish the foundation journey before starting agentic work. The two can move in parallel, with agentic value expanding as maturity improves.
From dashboards to orchestration: Why central governance matters
We also discussed the dashboard sprawl many consumer packaged goods (CPG) companies face today. Vibhor Mishra nailed one of the root causes: siloed dashboards often exist because a centralized platform doesn’t—so teams build what they need locally, using their own assumptions and definitions.
And that’s where agentic AI can become a forcing function—not by replacing dashboards overnight, but by creating a new layer above them: an orchestrator that can interpret signals, run scenarios, and recommend actions across levers.
But we were aligned on a key warning: if KPI definitions and return on investment (ROI) logic aren’t governed centrally, then agentic experiences will reproduce the same fragmentation—just faster. Vibhor Mishra emphasized that enterprise design choices (what must be standardized versus configurable) are as important as the technology itself.
The hidden value of agents: Speed to insight (not autonomy)
Perhaps the most provocative point in our discussion was the productivity shift.
Soudip Roy Chowdhury described how a decision request that typically takes a large team of analysts a week or two—to consolidate data, run analysis, iterate, and prepare a leadership-ready view—can become near-instant in an agentic model for information extraction and synthesis (not automated action).
This is where I think many leaders misjudge the adoption path. The near-term breakthrough isn’t “autonomous revenue management.” It’s radically faster cycles of decision intelligence—enabling business users to explore scenarios, pressure-test assumptions, and then bring analysts in to critique and deepen, rather than to assemble.
Human judgment remains central. Agents should recommend, suggest, and collaborate—not override.
Agentic AI is not magic, and it is not meant to replace the hard work of real Revenue Growth Management. What it actually does is cut through the noise so teams can focus on the judgment calls that matter. When you move from scattered dashboards to true decision intelligence, you do not get hype. You get clarity, speed and better choices
Marco Casalaina, Microsoft VP Product Core AI
Where Microsoft innovation fits: Horizontal platforms meet domain depth
A question I care deeply about—and asked directly—is how domain players stay aligned as Microsoft accelerates investments in AI platforms and agents.
Soudip Roy Chowdhury described a co-evolution dynamic: Microsoft provides horizontal capabilities, while domain solutions pressure-test them in real enterprise contexts—sending product feedback, such as benchmarking agent performance, and collaborating with Microsoft teams using tools like Microsoft Foundry and opensource components such as LangChain.
This is how modern enterprise innovation scales: platform, partner, and practitioner. Microsoft’s agentic investments can provide the secure foundation—identity, access, orchestration patterns, and governed data experiences—while domain partners bring the deep RGM decision journeys, ontologies, and workflow embedding required for adoption.
A practical takeaway: A readiness lens leaders can actually use
If you’re a CPG leader evaluating agentic RGM, here’s the simplest way I’d frame it:
- Confirm your system of record: Do you have account-level financial truth for trade and spend? Can you allocate funding cleanly across retailers and levers?
- Strengthen your system of intelligence: Can you standardize definitions, map data reliably, and operationalize models and simulations?
- Deploy agentic experiences where speed creates advantage: Start where faster insight loops deliver measurable wins: scenario exploration, cross-lever interpretation, anomaly detection, and recommendation support—with humans firmly in the loop.
- Add a deliberation layer that turns insights into action: Once data-driven hypotheses are formed, the agent convenes the right collaborators to pressure-test assumptions, build consensus, route decisions into the operational workflow, and continuously monitor outcomes—creating a living learning system that blends human and digital labor to execute complex work with end-to-end traceability.
This is how we move agentic AI in RGM from hype to durable value: decision intelligence grounded in business reality.
Explore solutions and more
- Explore how Microsoft AI for Retail helps consumer goods organizations modernize pricing, promotions, and decision intelligence.
- Learn how Microsoft embeds governance and accountability into AI systems through its Responsible AI practices.
1 McKinsey: Harnessing revenue growth management for sustainable success
