AI doesn't change the fundamentals of pricing

June 11, 2026

Head of Operational Excellence

Nordic B2B SaaS companies are quickly adding AI to their products. In our work with portfolio companies, we often see pricing lagging. The principles haven't changed, but AI makes it much harder to avoid the questions we've been putting off.

Here’s what we’re noticing and how we’re applying it to our pricing work moving forward.

The main principle stays the same, but AI raises the bar for what’s needed

Some people say AI completely changes pricing models. We don’t fully agree with that.

Value-based pricing, charging for what you actually deliver, not just access, has always been the right place to start. What’s different with AI is what creates value and how clearly you need to show it.

Ingvild Farstad, Head of Operational Excellence at Viking Growth

The core questions are the same as they've always been:

  • What value do we create? Increased revenue or cost savings?
  • What ROI can we demonstrate from incremental improvements to measurable impact?
  • How does the AI component specifically contribute to value creation for the customer?

AI just makes these questions tougher to ignore. If an AI agent does the work of three employees, you can’t price it like a single user license. The solution is the same as always: know the value you deliver and build a pricing model that matches it.

What we're seeing in the market

AI-driven products range from simple assistants to fully autonomous agents. Pricing usually shifts from charging per user, to per activity and then to actual outcomes. The closer your price matches real business results, the better the fit between price and value, and the more customers are willing to pay.

Three approaches we're seeing right now:

  • Consumption-based (per API call, per LLM token)
    This approach gives you clean margins and predictable costs. The challenge is that customers don’t think in tokens, so it mostly works for technical buyers right now.
  • Workflow/process-based (per completed task)
    The workflow-based model aligns with how work actually happens and offers a clearer value proposition. The challenge is higher cost variance. It works well when the task itself is what you’re delivering.
  • Outcome-based (per successful result)
    This approach gives the best match between price and value if you trust the AI’s performance and can measure the results. The challenge is that buyers take on more cost risk and face higher variance.


A hybrid model combining a fixed subscription with usage- or outcome-based pricing for AI features is a good place to start. It gives both sides predictability and lets you capture value from your results. Don’t forget to include the cost of running your AI products in your pricing.

Seven things we're taking into pricing work


These points matter whether you’re pricing AI or not, but AI makes them even more important.

  1. Balance new sales and expansion.
    Pricing and packaging matter for new customers and for future growth from your existing base. The model needs to be easy to sell today and easy to scale tomorrow.
  2. Meet customers where they are.
    Match the model to what customers are used to buying. That shifts over time, and you may need to educate your market, particularly when it comes to AI pricing.
  3. Remove friction.
    Make it easy to buy with predictable costs. Unpredictability kills deals, especially when the CFO is in the room.
  4. Review existing contracts.
    Understand your current commitments and renegotiate where needed.
  5. Make sure your product can actually support your pricing model.
    Can you make the changes needed to deliver what you’re charging for?
  6. Expect internal resistance.
    The loudest pushback on pricing changes often comes from inside the company. Involve sales and customer success early; they're the ones who have to sell it.
  7. Start testing price increases from day one.
    Set a price that feels slightly too high. Sharp customer feedback tells you what they actually value and where the ceiling is.

Learn while you sell

It’s a mistake to wait until you fully understand the value of your AI before charging for it. If you’re delivering real value but not pricing for it, your margins are shrinking without you noticing.

Treat your first pricing as a test you’ll improve over time, just like you would with product development. The Nordic companies that succeed aren’t always the ones with the perfect model. They’re the ones who start early, learn from sales conversations, and adjust as they go.

At Viking Growth, pricing is one of the first things we tackle with new portfolio companies. AI is making this work both more urgent and more interesting. We don’t have all the answers, but we’re learning quickly alongside the companies we invest in.

This piece was originally published in Norwegian by Shifter, interviewing Ingvild Farstad, Head of Operational Excellence at Viking Growth.

Latest News

More news