AI pricing puts software firms under margin pressure

Revenera has published its annual research on software monetisation trends, focusing on how established software companies and intelligent device manufacturers price AI features added to existing products.

The survey of 415 product and business leaders at global technology companies found that AI features are now common, but customer uptake remains uneven. According to the findings, 96% of respondents have added AI to their software products, while 48% said customer adoption was limited or low.

The report also highlights a gap between concerns about pricing pressure and the direct impact on customer seats. While 45% of respondents reported downward pricing pressure linked to AI, only 7% said they had experienced direct customer seat loss.

That suggests many software producers are facing tougher commercial discussions without yet seeing broad declines in user numbers. It also indicates that vendors are still testing how to package AI in ways customers will accept.

Pricing choices

The most common pricing approach was to place AI features in a higher-priced tier, cited by 40% of respondents. Close behind, 39% said they include AI at no extra cost, while 37% pointed to hybrid models combining subscription pricing with usage- or outcome-based charging.

Those choices reflect a broader preference for predictable billing over pure consumption pricing. Respondents were almost three times more likely to prioritise pricing predictability over near-term revenue maximisation, by 47% to 16%.

"Software buyers are balking at unpredictable AI spend that isn't clearly aligned to value. The next phase of AI adoption will hinge on getting the commercial foundations right," said Nicole Segerer, General Manager, Revenera.

Segerer also pointed to the operational demands behind AI pricing decisions. "With AI creating new revenue opportunities while putting pressure on profitability, software vendors must understand how customers use AI, where they see value, and what it costs to deliver. Usage data is critical to connecting those dots, enabling producers to turn usage insights into pricing strategies that encourage adoption, avoid surprise bills for customers, and protect margins for vendors," she said.

Margin pressure

The findings present a mixed picture for suppliers trying to turn AI additions into sustainable returns. Some (69%) of respondents said their AI offerings were driving revenue growth, and 52% reported improved profitability. Yet 41% said delivery costs were squeezing margins.

Higher AI delivery costs and rising cloud costs were identified as the main threats to growth. That tension helps explain why suppliers appear reluctant to rely too heavily on open-ended usage charging, even though AI services often carry variable operating costs.

The study also found that many companies lack the customer insight needed to set pricing with confidence. Respondents most often cited "data availability and insights" and "understanding customer value drivers" as the main barriers to aligning price with value, with both selected by 49%.

In practical terms, many companies are still handling usage data in a labour-intensive way. The research found that 41% still rely on manual processes or engineering effort to collect usage data, and only 27% use automated alerts to flag churn risk.

Slow change

The report suggests organisational structure may also be slowing pricing changes. More than half of respondents (55%), said it usually takes longer than six months to introduce a new monetisation model.

Responsibility for pricing and packaging decisions most often sits with product management (30%), or executive leadership (28%). Only 12% said dedicated pricing teams lead those efforts.

Analysts cited in the research said the findings challenge the view that software companies are moving rapidly and uniformly towards usage-based charging for AI. Instead, the data points to a more cautious market, where vendors are balancing demand for AI features with customer concerns about cost control.

"While industry conversations often suggest that everyone is moving rapidly to usage-based and AI-driven pricing, the reality is more nuanced. The Revenera Monetization Monitor: Software and AI Pricing Models and Strategies 2027 is valuable because it looks beyond the headlines and hype about monetization and pricing today. The data sheds light on most people's way of doing business, providing insights into how the 'quiet majority' of software and technology companies actually approach monetization," said Tiffany McCormick, Senior Director, AI Monetization, Pricing Strategies, & Business Models, IDC.

McCormick added: "Predictability, customer retention, and operational practicality remain just as important as innovation for many organizations evaluating how to monetize AI. The report complements the real-world evidence IDC is also finding, offering a clearer view of where pricing and monetization strategies actually stand, not just where the loudest voices say they're headed."

Another industry commentator quoted in the findings drew a similar conclusion about customer tolerance for variable charging.

"The data in the Revenera Monetization Monitor confirms a trend we're tracking across SaaS businesses: predictability matters more than maximizing every dollar of revenue. Buyers aren't rejecting usage-based pricing; they're rejecting an unbounded bill," said Rob Litterst, Co-Founder, PricingSaaS.

"Give customers visibility and real controls over consumption and you can charge for AI; skip that step and finance will cap you at whatever they can budget. The AI packaging that's working right now doesn't use the most aggressive metering: it's a predictable floor with a meter on top. Bundle an allowance into the tier so nobody has to build a business case to try the feature, then let usage do the expanding," he said.