The seat-based subscription that built modern software is colliding with a product that does not fit it. When the thing you are selling is an AI agent doing variable work, charging per human login stops making sense, and the industry has started repricing in public. The clearest signal is not a pricing-page tweak. It is an acquisition.
The seat does not fit the agent
As we reported, Salesforce has signed a definitive agreement to acquire m3ter, a metering and rating platform built for consumption-based monetization, and will fold its high-volume mediation, metering, and rating natively into Agentforce Revenue Management. m3ter ingests product-usage data in near real time, lets customers configure consumption-billing scenarios, and automates monetization across CRM, ERP, and quote-to-cash systems. Agentforce general manager Meredith Schmidt put the rationale plainly: AI shifts the landscape from traditional subscriptions to consumption-based models. The deal is expected to close in the second quarter of Salesforce’s fiscal 2027.
The tell is that Salesforce chose to buy this capability rather than build it. Metering at enterprise scale, accurately and in near real time, is hard infrastructure, and the company evidently decided it could not wait to assemble it internally while the pricing model shifts underneath its core business.
The arc: from licenses to seats to meters
Software pricing has reinvented itself roughly once a decade. Perpetual licenses gave way to the per-seat subscription, which became the dominant model precisely because it produced the smooth, predictable annual recurring revenue that investors learned to prize. Consumption pricing is not new either, it already governs the infrastructure layer, where the bill tracks usage because usage genuinely tracks value delivered. What is new is the model migrating up the stack into applications, where the unit of value is shifting from a person with a login to an agent completing a task. When the worker is software, counting seats measures the wrong thing.
It is not just Salesforce
The move reads less like a one-vendor bet when set against the wider market. SAP has publicly shifted toward AI consumption pricing as agents pressure the seat-based revenue model, according to trade coverage in ERP Today, and analysts tracking the space, including commentary aggregated by Seeking Alpha around HubSpot, describe AI monetization broadly migrating toward usage-based pricing. Bessemer Venture Partners has published an AI pricing and monetization playbook precisely because founders keep asking the same question. Even Salesforce’s own shares came under pressure earlier in 2026, with outlets such as Tech Times tying the slide to investor fear that AI agents threaten its seat-based revenue. When the incumbent most exposed to seat compression spends to own metering, the signal is that the repricing is not optional.
The metering problem is harder than it looks
Charging for usage sounds simple until you try to build it. You have to capture every relevant usage event accurately, attribute it to the right account and contract, deduplicate and rate it against whatever pricing scenario applies, and turn it into an invoice fast enough to be defensible, all across CRM, ERP, and quote-to-cash systems that were never designed to agree with each other. Get it wrong and you either undercharge and erode margin or overcharge and lose trust. That difficulty is exactly why Salesforce bought a specialist rather than bolting metering onto its existing billing, and why this becomes a revenue-operations problem rather than a finance afterthought.
The skeptic’s case
Consumption pricing is not a clean win. It transfers uncertainty to the buyer, who now faces a bill that varies month to month, and that anxiety can slow adoption or push customers toward competitors offering a predictable cap. It also complicates the recurring-revenue story that public-market investors reward, trading smooth ARR for a noisier top line. Many vendors will land on a hybrid, a platform fee plus usage, rather than pure metering, precisely to keep some predictability on both sides. The shift is real, but consumption is a dial, not a switch.
What it means for the revenue leader
The deepest change is that pricing becomes a data-infrastructure problem rather than a packaging decision, and that work lands squarely on revenue operations. The team that could once express a price in a spreadsheet now needs a metering pipeline, and the vendor that can ingest events and produce a defensible invoice in near real time controls how AI products get monetized. Revenue leaders will also have to explain a less predictable top line while building the instrumentation to make it legible. The teams still running purely on flat-seat assumptions will struggle to price agentic products at all, and will discover it at renewal, when a customer asks why they are paying for 200 seats to run a handful of agents.
What to evaluate now
Ask whether you can meter product usage in near real time today, and if not, treat that as a revenue-operations priority. Examine how exposed your own model is to seat compression as agents replace logins. Decide deliberately between building metering and buying it, knowing the largest vendor in the category just chose to buy. And start tying price to outcomes where you can defend the measurement, because pay for what the agent did is the question your customers are about to start asking out loud.
The repricing will not happen all at once, and the vendors that handle it best will treat it as a transition to manage rather than a switch to flip. But the direction is set. In an agent-driven market, the company that can prove what was used, and bill for it cleanly, holds the pricing power, and that capability now sits at the center of the revenue stack rather than the edge of finance.
