How Much Revenue Is Your AI Managing?
The first phase of enterprise AI has been measured largely through adoption. How many employees are using it? How much time are they saving? How many analyses can be automated, workflows accelerated, or questions answered without the manual effort that previously sat behind them? These are sensible measures for an emerging technology, particularly when most AI is still being introduced as another tool employees use to perform their existing jobs more effectively.
They may not be the measures that matter for long.
As AI becomes embedded more deeply into commercial operations, the more important question will be how much responsibility an intelligent system can assume for continuously improving the business around it. Not replacing the people responsible for pricing, forecasting, promotions, assortment, customer strategy, or retailer relationships, but giving those people a system that is always monitoring performance, identifying changes, evaluating alternatives, recommending action, and learning from the decisions they ultimately make.
We think of the economic activity operating within that model as Revenue Under Management.
Revenue Under Management is not a measure of how much revenue AI controls autonomously. It is the revenue for which a Commercial Intelligence System maintains persistent intelligence: continuously understanding what is changing, determining where intervention may be required, evaluating what could happen next, and helping commercial teams make better decisions as conditions evolve.
That distinction matters. The ambition for enterprise AI should not be to remove people from commercial decision-making. It should be to remove the enormous amount of friction that prevents talented commercial teams from applying their judgment where it creates the most value.
Software Has Always Waited for Us
Traditional enterprise software is fundamentally reactive. An analyst opens a forecasting application because it is time to update the forecast. An RGM team runs pricing scenarios because a pricing decision needs to be made. A category team reviews assortment because a retailer reset is approaching. Executives open dashboards because they want to understand why performance changed.
The software can be extraordinarily sophisticated once someone engages with it, but responsibility for identifying the problem, initiating the analysis, connecting the implications across functions, and determining what happens next still largely sits with people.
This creates a substantial amount of hidden work. Commercial teams spend time monitoring dashboards for changes worth investigating, reconciling different versions of performance, preparing recurring analyses, determining which exceptions matter, rebuilding scenarios when assumptions change, and coordinating conclusions across teams. None of these activities is inherently unimportant, but together they consume an enormous amount of organizational capacity before judgment is ever applied to the commercial decision itself.
A Commercial Intelligence System changes that relationship because it does not need to wait for someone to ask the next question. If it maintains a persistent understanding of the business, it can continuously evaluate actual performance against expectations, identify where assumptions are changing, determine which developments are commercially meaningful, and bring the right decision to the right people with the relevant context already assembled.
The system becomes more active without requiring the organization to become less human.
Shifting from AI Adoption to AI Responsibility
This changes how we should think about AI adoption.
A commercial organization could give every employee access to AI and still operate essentially the same way it does today. People would complete individual tasks faster, analyses might take hours instead of days, and productivity would improve, but the underlying operating model would remain dependent on people continuously moving information and decisions through the organization.
The larger opportunity is not simply broader AI usage. It is giving intelligence persistent responsibility for defined commercial problems.
Consider a portfolio representing $500 million of revenue across products and retailers. Today, teams may review different portions of that portfolio at different cadences, with attention determined by planning calendars, retailer meetings, reporting cycles, and whatever issue is most urgent at a particular moment. A Commercial Intelligence System can maintain a continuous view across that same portfolio, evaluating whether demand is moving away from forecast, whether a competitive price change creates a new opportunity, whether promotional performance is deteriorating, whether an assortment decision is producing unexpected transfer, or whether a previously recommended action should be reconsidered as conditions change.
The system does not need authority to make every decision for that model to be transformative. Its responsibility is to make sure the commercial organization knows when a decision should be made, understands the alternatives and expected consequences, and can act before the opportunity disappears.
That is the shift from measuring users to measuring responsibility.
Revenue Under Management
Enterprise software has traditionally been priced and evaluated through seats, modules, transactions, workflows, or consumption because those measures reflect how people use software. As intelligence assumes a more persistent role in commercial operations, a more meaningful measure may be the economic activity for which the system is responsible.
We call that Revenue Under Management.
A category representing $100 million of annual sales could be under management even though its most consequential pricing and assortment decisions remain firmly in human hands. The system’s role is to maintain continuous intelligence around that revenue: monitoring the environment, updating expectations, identifying material deviations, evaluating possible responses, surfacing recommendations, and then learning from whatever decision is ultimately made.
As confidence grows, the nature of that responsibility can evolve. Some activities may remain advisory, particularly where strategy, customer relationships, negotiation, or significant economic tradeoffs are involved. Others can become increasingly orchestrated, with the system coordinating approved actions across workflows and applications. Routine decisions with well-understood parameters may eventually operate within predefined guardrails without requiring the same level of manual intervention.
There does not need to be a single destination for every decision. The objective is to establish the right division of labor.
People should spend their time where human judgment creates advantage. Intelligence should assume more of the continuous analytical burden required to make that judgment effective.
Why AI Increases the Value of Commercial Expertise
There is an understandable tendency to frame advances in AI around what people will no longer need to do. In commercial CPG, that framing misses where much of the opportunity actually lies.
Pricing is not simply a mathematical optimization problem. Retailer relationships matter. Brand strategy matters. Competitive context, organizational priorities, negotiation dynamics, consumer perception, and management’s appetite for different tradeoffs all matter. The same is true across promotions, assortment, forecasting, and customer planning. Better models can dramatically improve these decisions without eliminating the need for people who understand the business and are accountable for its direction.
What should change is how much of those people’s capacity is consumed simply getting to the decision.
A Commercial Intelligence System can continuously monitor the portfolio rather than requiring teams to search for problems. It can maintain forecasts rather than periodically rebuilding them, evaluate thousands of possible commercial changes rather than waiting for an analyst to construct individual scenarios, and preserve context across functions rather than requiring people to reconstruct it in meetings. As decisions are made and executed, the system can then observe the resulting outcomes and incorporate that experience into the next recommendation.
The commercial leader’s role moves upward rather than outward. Less time is spent operating the analytical machinery around the business, while more time becomes available for deciding what the organization is trying to accomplish, challenging recommendations, applying context the system cannot fully observe, managing customers, and making the strategic tradeoffs that ultimately define the business.
The system does not replace commercial expertise. It gives that expertise leverage.
A Continuously Intelligent Business
The concept of Revenue Under Management becomes possible because of the evolution described throughout this series.
A Commercial Intelligence System first connects decisions that point solutions have historically optimized independently. It creates new proprietary intelligence from the organization’s data, connects recommendations to execution, observes what happens in market, and learns from the resulting outcomes. Once that loop exists, intelligence no longer needs to appear only when someone opens an application. It can remain continuously engaged with the economic activity it is responsible for supporting.
That creates a fundamentally different commercial operating model.
A traditional application can tell a team how to improve a forecast when the team asks it to run one. A Commercial Intelligence System should know that the forecast has become less reliable before the next planning meeting occurs, understand why, determine which commercial decisions are affected, and help the organization decide whether action is warranted.
The same principle applies across pricing, promotions, assortment, customer planning, and execution. Instead of periodically applying intelligence to individual decisions, the organization places increasing portions of its revenue inside a continuously intelligent environment.
The ambition is not autonomous revenue – it is continuously managed revenue, with human and machine intelligence each doing what they do best.
A New Measure of Commercial Intelligence
The value of a Commercial Intelligence System should ultimately be measured by what it changes in the business. How quickly can an organization recognize that a decision needs to be made? How much intelligence can it bring to that decision before the opportunity passes? How effectively can it understand the implications across pricing, demand, promotions, assortment, retailers, and financial outcomes? And, ultimately, does that lead to better commercial results?
Revenue Under Management provides a different way to think about that opportunity. As more of the business operates within a continuously intelligent environment, commercial teams no longer need to wait for the next planning cycle, performance review, or analytical request to discover that conditions have changed. The system can continuously evaluate the business, identify where intervention matters, and bring forward the decisions most likely to affect performance with the context and alternatives already understood.
The result is not decision-making without people. It is better decision-making with far less friction around them. Commercial teams can respond sooner, evaluate more possibilities, understand tradeoffs more completely, and apply their judgment to a higher-quality set of choices. As those decisions are executed and their outcomes return to the system, the intelligence behind the next decision improves as well.
At Insite AI, that is how we think about Revenue Under Management. The goal is to place increasing portions of the commercial business inside an operating model where intelligence is persistent, decisions happen faster, and every outcome contributes to what the organization knows next.
That changes the question commercial leaders should be asking.
Not simply: How much data do we have, or how sophisticated are our tools?
But increasingly: How much of our revenue is being continuously informed by better intelligence?
Because the real measure of commercial intelligence isn’t the technology behind it, it’s the speed and quality of the decisions it enables and the outcomes those decisions produce.
