
Usage-Based Pricing: Why Enterprise Software Contracts Need Renegotiating Before 2028
IDC predicts that 70% of enterprise software vendors will shift from fixed licensing to usage-based models by 2028. For mid-sized businesses, this means today's contracts could become costly and opaque — and preparation needs to start now.
AI agents are handling tasks that previously required twenty user licenses. This doesn't just change workflows — it calls into question the fundamental logic of per-seat licensing. Why should a mid-sized company still pay based on the number of desks when a single software agent can do the work of ten employees?
The software industry's response: usage-based pricing. And this shift is arriving faster than most businesses expect.
What IDC and Forrester predict for 2028
IDC FutureScape 2026 (as analyzed in CIO.com), puts it plainly: 'By 2028, pure seat-based pricing will be obsolete — AI agents replace repetitive tasks with digital labor, and 70% of vendors are forced to refactor their value proposition into new models.' Salesforce and Workday have already announced this shift in investor calls: this is not a niche trend but an industry-wide move.
For 2026, Forrester describes concrete transition models: so-called Agentic Enterprise License Agreements (AELAs) — flat-fee contracts covering unlimited AI agent usage. Salesforce markets this as 'All You Can Eat with Agentforce.' For buyers, this initially sounds attractive. The risk: companies with today's classic seat-based contracts will have to negotiate under entirely different conditions in two years.
What usage-based pricing means for your budget
Usage-based pricing is not inherently worse but it fundamentally changes the risk profile for buyers. Until now, software budgets were predictable: 50 licenses × €80 = €4,000 per month. Going forward, the bill depends on how intensively the system is used, how many AI queries are made, and which thresholds the vendor defines. Forrester already warns of a dynamic resembling uncontrolled cloud spending: variable costs that are difficult to budget during peak usage. Proactive IT consulting addresses exactly this challenge.
- Vendors use contract transitions to improve margins — without active preparation, existing customers often end up paying more in total.
- What counts as 'usage' is defined by the vendor: a login, an API call, or a completed workflow can all be valued very differently.
- The deeper AI agents are integrated into processes, the harder it becomes to switch vendors — even when terms become unfavorable.
- Budget overruns in consumption-based models are difficult to detect and manage without active monitoring.
When an AI agent replaces a human task, customers rightfully expect to pay based on outcomes — not on the number of logins.
Four measures mid-sized businesses should take now
The 2028 forecasts are not a cause for panic but a clear signal for proactive action. With a structured technology assessment, your negotiating position can be significantly strengthened before vendors reset the terms:
- Software inventory: Which licenses are actually being used? Barely-used licenses are direct negotiating leverage and reveal where consolidation potential exists.
- Review contract terms: When do current contracts expire? Negotiations should begin 12–18 months before renewal — not after the vendor has already set new terms.
- Reassess build vs. buy: When standard software becomes more expensive and complex, comparing against custom software is worthwhile — especially for core processes with specific requirements that no standard vendor covers without customization effort.
- Include AI strategy in procurement: Anyone planning AI integration and automation should examine early whether license models reflect AI usage fairly — or whether the vendor's new model multiplies costs.
The shift to usage-based pricing is inevitable. Companies that analyze their software contracts and strategy through this lens today will have significantly stronger negotiating positions in 2028 than those caught off-guard by the changes.