Rethinking Enterprise Software Pricing Models For The AI Era

The last great disruption to enterprise software pricing models occurred in the early 2010s, as B2B software vendors shifted from one-time payments for on-premise deployments to subscription-based models for web apps utilizing cloud compute. The change brought major market implications and still continues to plague many vendors. Providers were forced to move customers through painful migrations, sunset aging solutions and re-focus on customer retention.

In 2025, enterprise SaaS vendors across all market segments are increasing AI feature integration and, in some instances, acquiring AI-native software to accelerate the transition (see ServiceNow’s acquisition of Moveworks as a recent high-profile example). AI’s relative immaturity is creating major uncertainty around commercial models. Vendors are grappling with how to price AI features; whether entirely new pricing models should be deployed; and managing variable inference compute costs –all while encouraging AI adoption. Here are the key learnings so far:

• Enterprise SaaS vendors’ AI pricing is very much in an experimentation stage.
  There is no hard and fast rule for SaaS AI pricing strategies in 2025. Factors influencing decisions include AI maturity of the market segment, AI use case commoditization, and whether the use case can be considered a ‘feature’ or if it’s more of a distinct ‘product’. However, some patterns are emerging. For instance, vendors operating in markets with comparatively immature AI uptake, such as EHS software or sustainability reporting software, are generally looking to encourage AI adoption by baking AI features into the standard pricing (rather than charging a premium). To increase deal flow for AI-first tools, vendors like C3 AI are focusing on GenAI pilot projects, with the aim of getting PoCs in front of big clients before scaling up. Approaches remain divergent and we expect to see lots of trial and error as vendors figure out what works.

   

• Vendors should stay attune to AI feature commoditization.
  As AI feature development powers onwards, Verdantix expects certain use cases to quickly become commoditized, raising the baseline buyer expectation. Key examples of this are enterprise data semantic search and integrated virtual assistants. At present, many of the largest SaaS businesses are charging separately for AI ‘products’ – consider Salesforce Einstein and Microsoft Copilot. In the long run, some capabilities will transition from differentiators to a core need as part of a minimum viable product, thus pushing pricing to vary based on the sophistication of AI features rather than on-off toggles. GenAI marketing platform Jasper withholds features like no-code AI app building and variations of marketing tone types or ‘voices’, from all users except its most premium tier.

   

• Vendors are breaking the mould with value- or outcome-based pricing.
  Existing enterprise SaaS pricing models can be highly opaque, but generally comprise some combination of number of users, functional breadth, firm size and functional complexity. As AI agents’ ability to autonomously complete tasks increases, user numbers no longer correlate nicely to software value. Consider a world where one user controls a hoard of autonomous agents, replicating the work effort of hundreds of end users. There are already excellent examples of B2B software firms with outcome-based pricing models – meaning clients only pay for specified outcomes of a certain quality. Sierra, for example, is an AI-driven customer support vendor that only charges clients for customer tickets resolved by its AI agents, rather than subscription or usage-based pricing.
  
  

In summary, AI feature pricing and underlying pricing models are in a state of flux, with 2025 being a time for experimentation and longer-term pricing strategy planning. For more insights relating to AI-fuelled disruption, visit the Verdantix website. For guidance on marketing strategies for AI, check out our latest webinar: Keeping Your AI Strategy On Track: What Product & Marketing Teams Need To Know.