In the monolithic SaaS pricing playbook, differentiation from one company to the next lay in product features and positioning. We’re entering a phase where the business model itself is a source of differentiation. The pricing model becomes a stronger way of differentiating a company to a customer segment, reducing competition with category peers. Why? Because AI products are delivering value in a variety of new ways: co-pilots, agents and service-as-software.

While the seemingly magical AI demonstrations grab the headlines, this shift isn’t just about new product features — it’s affecting go-to-market strategy, brand-building, customer segmentation, team evolution — and  long-term business model design. Pricing isn’t a color-in-the-boxes afterthought. It’s a strategic weapon.

Pricing used to be predictable. Not anymore.

Until a few years ago, early-stage SaaS companies rarely worried about pricing beyond ‘basic, pro, enterprise.’ You would launch with a three-tier model, tack on some integration upsells, make sure your sales team knew how to customise quotes for the enterprise segment, and get back to shipping features. In a world dominated by seat-based pricing models, this worked— especially if your core product centered around automating workflows. Most SaaS products delivered value by taking a process and recreating it in a UI. By users clicking and typing, a company collected its data and workers became more efficient, generating ROI by increasing revenue or saving costs. Value scaled similarly across different software categories.

Then came usage-based pricing, popularised by infrastructure players like Snowflake and Datadog. That was the first crack in the wall of the license-based pricing metric which reigned synonymous with SaaS for a long time. ROI for a SaaS product no longer always scaled neatly with more users.

With AI, this wall is poised to crumble. Co-pilots turbocharge existing paradigms. Agentic AI is commonly positioned to replace human labor, making seat counts obsolete. Service-as-software models may or may not break apart from the billable hour or fixed-cost models prevalent in the professional services industry.

In other words: good luck predicting the dominant future business model in software.

New companies have the advantage

Startups have a golden opportunity here. With no existing user base, they can afford to experiment aggressively on the journey to product-market fit and beyond. Should you price based on usage? Per outcome? Per agent? A base platform fee plus per-action charges? Or will seat-based pricing with modifications work the best? It’s a really open question, and as a newcomer you get the luxury to test without risking customer ire. Find out how your target customer segment best perceives value - and build your product, business model and Go-To-Market around that. 

For companies with large existing customer bases, it’s trickier. Roll out radical product upgrades that allow customers to reap the same rewards with far fewer seats and you potentially cannibalise the revenue base. But standing still is just as dangerous. AI-native challengers see you as a target — and their pricing model might become the one your customers now expect.

We can see this dilemma playing out in the real world. New companies are likely to experiment with outcome-based pricing models. Established incumbents roll out AI-product features, but may carve them out as an upsell for the enterprise tier. Early-stage investors herald outcome-based pricing as the future; later-stage investors are a bit more skeptical.

Moving past the “AI is special” moment

Right now, many companies are pricing AI features as add-ons or premium upsells within their existing pricing models. By and large, this has been a response to high LLM model inference costs. But that’s a short-term game. As foundational model costs continue dropping precipitously and basic AI features become table stakes, those features will commoditise — and with them, their margins.

Similarly, in a world moving quickly past AI as a ‘fun add-on’, marketing the AI unto itself is unlikely to persist long into the future, except on the websites of poorly positioned products. Buyers are increasingly clear that while AI might be a ‘must-have’ in certain product categories, their core question remains the same: what’s the differentiated value (from all the other AI-powered products) and what’s the ROI?

Suddenly, we’re right back to where we started. Pricing power stems from a product with differentiated value and clear positioning. That power may manifest in a variety of pricing models

A messy mix, not one model to rule them all

Will every sector move at the same pace? Absolutely not. Some categories — like marketing tech — are at the forefront, with marketers rapidly experimenting with tools to personalise and automate lead generation. Some vertical software categories are just now moving from spreadsheets to software. Digitising workflows is still a key source of value, and seat-based pricing models featuring co-pilots may suffice. Customers’ desire for established, predictable pricing models may be very different.

What’s clear is that we’re not heading for one dominant model, but a mix. Some products will remain classic subscriptions. Others will charge based on consumption. Others on agents. Others on outcomes. And many on a hybrid model.

With a larger menu of pricing metrics, what does that mean for companies? First, they should spend more time on pricing design. Second, they need to be more willing to experiment. And finally, they need to think from first principles rather than copying larger SaaS luminaries as each sector is likely to develop at different rates.

AI forces software founders to ask the hard questions:

  • What kind of value are we really delivering?
  • How does that value really scale?
  • How can we make pricing a competitive differentiator to win our customer segment?

The future: a pricing practitioner’s dream

Ultimately, the companies that win won’t be the ones that add AI features the fastest. They’ll be the ones that launch value-driving AI features — and design complementary monetisation models. Pricing has always been a signal for how a company conceives value and the type of customer it looks to attract. The company’s ability to align price with perceived value better than incumbents is a key source of competitive advantage. In the AI age, pricing is no longer just a lever — it’s a moat.

Ingrid Bonde Åkerlind is Principal at Oxx