According to ZDNet, software pricing is heading for a massive shakeup, moving from traditional per-seat licenses to outcome-based models where companies pay for actual results. This shift, predicted to be bigger than the move from disks to SaaS, requires vendors and users to agree on specific success metrics, like Zendesk’s model of charging for fully automated customer issue resolutions. Analysis from West Monroe forecasts the end of per-seat licensing, noting that 12% to 15% of enterprise IT budgets now go to AI, favoring AI-native providers. Furthermore, 80% of engineers will need to upskill for AI-driven roles in the coming year as AI reshapes software development, leading to leaner engineering teams that deliver more output.
Pay for what works, not just what you get
Here’s the thing: paying for outcomes sounds fantastic in theory. You only cough up cash when the software actually does the job you hired it for. But man, is that going to be messy in practice. Defining those success metrics? That’s where the real negotiations begin. It’s easy for a vendor like Zendesk to say “pay per fully resolved ticket,” but what about more complex enterprise software where “success” is fuzzier? This forces a much deeper, more collaborative relationship with your vendors. It’s not a one-time purchase anymore; it’s a continuous partnership built on shared data and, frankly, a lot of trust. And as West Monroe points out, this breaks the old assumption that software costs are fixed. Your CFO is going to have a whole new set of headaches trying to forecast a bill that fluctuates with your business activity and AI’s performance.
AI reshapes the builders, too
This isn’t just about how you buy software; it’s about how it’s made. The report’s note that 80% of engineers need to upskill is huge. We’re talking about a fundamental change in the job. AI is automating the routine stuff—code generation, testing—which theoretically means leaner teams. But does that mean mass layoffs? Probably not anytime soon, as the experts emphasize. Think about it: if AI handles the grunt work, what’s left for humans? The complex problem-solving, the innovation, the deep customer engagement that machines can’t fake. The skill set just shifts higher up the value chain. The risk is a brutal talent divide between those who can work with AI and those who can’t.
How to get ready
So, what does an IT leader do now? First, you start vetting vendors differently. You’re not just looking for a product; you’re looking for a partner in “continuous value creation.” You need transparency into AI performance and usage. You need vendors who will help you forecast spend and benchmark success. Basically, you need a business ally, not a software salesman. Second, look at your own internal processes and contracts, especially for outsourcing. If you’re still paying based on labor hours and headcount, your incentives are totally misaligned in an AI world. You want contracts with shared savings models, so both you and your partner benefit from AI-driven efficiency gains. It’s a whole new way of thinking about value.
The bottom line
This feels inevitable, right? As AI agents start doing more of the actual work, paying for “seats” or “users” makes less and less sense. You’re paying for the intelligence and the outcome, not the login credential. The companies that figure out these new partnerships and metrics first will have a real advantage. For everyone else, there’s going to be a painful period of adjustment where finance, product, and engineering teams have to talk to each other—a lot—because cost and usage are now directly linked. The era of writing a check and forgetting about it is over. The software bill just became your most important business performance report.
