Two Deals, One Pattern
Asana closed its acquisition of StackAI, a no-code AI agent orchestration platform, for approximately $75 million on May 28, 2026 (TechCrunch, 2026). Four days later, Salesforce announced a definitive agreement to acquire Contentful for roughly $1.5 billion in cash, net of shares Salesforce already held, with the deal expected to close in the third quarter of Salesforce's fiscal 2027 (Salesforce, 2026).
The two deals look unrelated on the surface. One is a mid-sized work management vendor buying an AI workflow startup. The other is Salesforce adding a headless content platform to feed its Agentforce and Data 360 products. But both moves share a mechanism worth naming plainly: neither company is folding the acquired capability into its existing per-seat plan. Both are set up to bill for it separately, on usage.
Asana's own pricing page shows how this works in practice. AI Studio, the workflow builder StackAI now extends, includes a base credit allotment on paid plans, then charges $150 a month for an additional 100,000 credits under AI Studio Plus, or moves to an annual contract with 5,000,000 credits per quarter under AI Studio Pro, with overage top-ups priced at $499 for 2.5 million extra credits (Asana pricing, as reported by usecarly.com and agiled.app, 2026). None of that is a per-seat number. It is a metered account on top of the seat count a company already pays for.
Salesforce has not yet published Contentful-specific pricing since the deal has not closed, so we are not asserting one. What Salesforce has stated is the intent: connect Contentful's content APIs to Data 360 and Agentforce so that AI-assembled content becomes a chargeable capability inside those products (Salesforce, 2026). Agentforce itself already bills on a consumption basis distinct from Salesforce's core per-seat CRM contracts. Attaching a newly acquired content layer to that consumption product, rather than to the flat-rate CRM seat, is a pricing decision, not just a product one.
The Market Data Backs It Up
This is not two data points forced into a trend. Independent market data shows the same shift happening broadly. PricingSaaS, which tracks pricing changes across 500 major software and AI companies, found that 79 of those companies now offer a credit-based pricing model, up from 35 at the end of 2024, a 126% increase in roughly eighteen months. New entrants to the credit-model column in 2026 include Figma, HubSpot and Salesforce itself (PricingSaaS, 2026).
HubSpot is a useful case because it moved in the open. Effective April 14, 2026, HubSpot restructured its Breeze AI agents onto outcome-based pricing: $0.50 per resolved conversation for the Customer Agent, $1 per lead recommended for outreach for the Prospecting Agent, and $0.10 per answer for the Data Agent (CMSWire, 2026). Those agents did not exist as a line item on a HubSpot invoice two years ago. Now they are billed per action, independent of how many seats a customer has already purchased.
The buyer-side data confirms this is landing as a surprise more often than not. Zylo's 2026 SaaS Management Index, published January 29, 2026, found that 78% of IT leaders had already been hit with unexpected charges tied to consumption or AI-feature pricing in the prior year, and that 61% had been forced to cut a planned project because of unplanned SaaS cost increases (Zylo, 2026). Zylo sells SaaS spend management software, so it has a commercial interest in this finding being true. The direction is nonetheless consistent with the vendor-side pricing data above, which comes from an independent tracker with no stake in either outcome.
What Vendors Are Actually Buying
Read the acquisitions and the pricing data together and the logic is straightforward. An AI feature built in-house usually ships as part of an existing tier, because it was budgeted as ordinary product development against a plan that already has a price. An AI feature bought through acquisition arrives with its own cost basis, its own team, and often its own existing pricing model from the company that built it. That gives the acquirer a clean argument, internally and to the market, for launching it as a new line rather than absorbing it into a flat rate that was set before the acquisition happened.
StackAI's original standalone product had already sold on usage before Asana bought it. Contentful's own content-delivery pricing has consumption components predating the Salesforce deal. In both cases, the acquirer inherited a metered pricing model along with the technology, and kept it, rather than replacing it with a seat-based number that would have required the buyer to eat the AI infrastructure cost inside a fixed contract.
The Buyer Impact
For procurement and IT teams, the practical effect is that an acquisition announcement in a vendor's stack is now a pricing event, even when no press release uses the word "price." A company running Asana should treat the StackAI integration into AI Studio as a future line item to budget for, not a free upgrade. A company running Salesforce should ask, before the Contentful deal closes, whether Contentful-powered content generation will be metered under the existing Agentforce consumption pool or billed as a new SKU, and get that answer in writing before renewal.
The broader lesson from the PricingSaaS and Zylo data is that this pattern is now common enough to negotiate around rather than accept as a surprise. Before signing or renewing any contract that includes AI features, ask the vendor for the credit or unit cost behind the feature, not just the sticker price of the plan. Ask what happens at the overage threshold. Ask whether a future acquisition-driven feature would be added to the current plan or billed separately, and try to get a cap or a most-favored-pricing clause written into the contract rather than discovering the answer at renewal.
What We Think
We think it is too early to call this the end of per-seat pricing outright. Most enterprise SaaS contracts we can point to still price the base product on seats and add consumption only for the AI layer sitting on top. But the direction of the last few months is clear enough to plan around: when a vendor buys an AI company, assume the acquired capability shows up as a metered add-on, not a free feature, and ask about the unit economics before you need to.