Microsoft is moving parts of its AI portfolio beyond pure per-seat pricing by pairing user licenses with charges tied to usage. The hybrid model keeps recurring seat fees in place while adding a variable billing unit as customers consume more AI capacity.

That model is beginning to show up in Microsoft’s financial results. Dynamics 365 Customer Service credit consumption rose fourfold quarter over quarter, while GitHub Copilot generated what Microsoft called “significant consumption revenue” after its June billing change.

Microsoft also reported more than 30 million paid Microsoft 365 Copilot seats, with net additions more than doubling from the previous quarter, and said GitHub continued adding Business and Enterprise seats.

Forrester warns that customers should brace for bigger software bills next year as software and AI vendors raise prices and pile on usage charges.

Working from a survey of more than 2,600 business and technology decision-makers, the tech research company said software budgets were expected to rise “as vendors increase prices or add usage charges to pass their AI costs to customers.”

In the last six months, Anthropic, OpenAI, and GitHub have shifted some services away from flat-rate subscriptions toward usage-based billing, prompting cost concerns among users. Forrester added Microsoft to the list, citing its recent launch of the premium E7 license, which bolts M365 Copilot, Agent 365, and security tools onto E5.

Last year, consultants Bain & Company estimated that the build cost for AI datacenters would hit $2 trillion by 2030.

Forrester said that AI would drive increases in data and software spending, with 80 percent of decision-makers expecting those budgets to rise.

Sharyn Leaver, chief research officer at Forrester, said: “The organisations that outperform in 2027 won’t be those that spend the most on AI. They’ll be the ones that invest in the foundations that make AI effective: trusted data, strong governance, organisational readiness, and the ability to continuously adapt as technology and customer behaviour evolve.”

Forrester also found that personnel costs have yet to fall, despite the “AI washing of layoffs” in the tech industry.

“While several tech giants, including Oracle, Microsoft, and Meta, have announced significant layoffs in recent months, IT staffing spend has not declined in recent years,” the report said.

Staffing accounted for 35 percent of IT budgets in 2025. For 2027, 67 percent of tech decision-makers expected to increase their staffing budget, while 23 percent said it would stay flat, and 10 percent expected it to decline. “The AI washing of layoffs will continue as vendors trim for financial and restructuring reasons. Guard against inflated promises that AI can replace employees across the board. Staffing for data/analytics-specific roles is expected to rise, with 68 percent of data technology decision-makers expecting this budget item to increase,” the report said.

Forrester said that organisations should adapt their FinOps practices to help manage the unpredictable costs associated with AI. “Traditional FinOps wasn’t built for token-based, usage-driven AI costs, but that team is certainly best positioned to build these new capabilities and must make this leap in 2027. Fund runtime cost controls such as model routing, semantic caching, and usage guardrails to prevent runaway spend,” the report recommended.

In July, KPMG research found that nearly a third of corporate leaders reported difficulty understanding and controlling operating costs when implementing business AI at scale. “As usage-based pricing models become more common, many organisations are still building the capabilities required to forecast, monitor, and manage AI spending effectively,” the consultancy said.

Software companies are leaving money on the table because their core financial systems haven’t kept pace with the way they sell pay-per-use services, which often now incorporate AI capabilities.

According to a survey of 350 execs who work for software vendors, conducted by PwC UK and billing infrastructure biz m3ter, 44 percent of UK business leaders say they have trouble measuring the consumption of usage-based software.

PwC and m3ter claim this measurement blind spot leads to revenue leakage – the gap between the value of what the company sells and the amount actually billed.

This can take the form of failure to capture usage data, like a subscription overage fee that isn’t tracked. Or it might reflect out-of-date pricing information, or billing calculation errors.

M3ter claims that between 4 and 7 percent of annual recurring revenue is potentially at risk from unsophisticated bookkeeping.

AI services can compound the problem, with the pricing of AI software and services often opaque and difficult to predict.

With regard to the way that AI complicates billing for companies that use it in their products, the research points to PwC’s 2026 CEO survey that found “only 30 percent of companies surveyed reported increased revenue from AI in the last 12 months.”

“The monetisation opportunity exists, but it is clear that converting AI capability into sustainable, auditable revenue remains a challenge,” the PwC/m3ter report states.

“A useful comparison is with the cloud infrastructure sector,” Griffin Parry, CEO and co-founder of m3ter, said in an email. “They had similar challenges and over time they developed greater transparency, via both better pricing design and better accessibility of usage and spend data.”

But the survey results aren’t solely reflective of black-box AI billing – sometimes it’s just inefficient business processes like tracking multiple product lines that combine seat-based and usage-based billing in a spreadsheet.

The PwC/m3ter survey found that 87 percent of respondents reported that their billing systems and their ERP or general ledger systems were not integrated. And 48 percent said there was no integration between billing and CRM systems.

Parry argues that transparency is a key theme for corporate customers, in terms of pricing design and the accessibility of usage and billing data.

“Good usage pricing should be easy for customers to understand, easy for them to predict (i.e. they can estimate how much they’d pay in real world scenarios), and with the variable element attached to a good proxy for value (i.e., so that if they use more of your service, they’ll associate that with being successful),” said Parry. “That’s one way of interpreting transparency, the opposite of opaque.”

Parry went on to say that customers won’t be satisfied with one monthly invoice that summarises usage-based spending.

“Instead, they’ll want to be able to access information about their usage and spend on a near-real-time basis, and for this data to be detailed – i.e., they want access to a usage and billing dashboard, as, say, AWS customers do,” he explained. “That’s the other way of interpreting transparency, access to the data that drives bills.”