How acquirers read AI risk
The new diligence question
For a decade, software buyers asked whether a product would keep its customers. In 2026 they also ask whether an AI agent could do the product’s job. The answer now shapes the multiple as much as growth or retention, and it is why two companies with identical ARR can receive very different offers.
Where exposure sits
| Product type | Exposure | Why |
|---|---|---|
| System of record (ERP, core banking, EHR) | Low | Agents depend on the data it holds |
| Infrastructure and developer platforms | Low | Usage rises as AI workloads grow |
| Vertical workflow with deep integrations | Moderate | Defensible if data and compliance are owned |
| Horizontal seat-based productivity tools | High | Fewer seats as tasks are automated |
| Content, reporting and code generation | Severe | Directly substitutable by models |
Moving pricing before a sale
Shifting from seats to usage or platform pricing takes two to four renewal cycles to show in the numbers. Founders planning an exit in the next two years should start now, with new customers first, so buyers can see the new model working in the cohort data.
Frequently asked questions
- Will AI agents replace SaaS?
- Not evenly. Products that are thin interfaces over a workflow, priced per seat and easy to switch away from are the most exposed. Systems of record with proprietary data, deep integrations and usage-based pricing tend to gain, because AI agents need to read from and write to them.
- How are acquirers pricing AI risk in 2026?
- Buyers now run an AI review alongside commercial diligence. Companies with monetised AI features and proprietary data can earn a premium over sector multiples; seat-based tools with no AI roadmap are increasingly priced at a discount, or bought for their customer base rather than their product.
- Why is per-seat pricing a risk?
- If AI lets a customer do the same work with fewer people, a per-seat contract shrinks at renewal even when the customer is happy. Usage, outcome or platform pricing grows with throughput instead, so it benefits from the same trend.
- What makes a software company AI resilient?
- Proprietary or network data, being the system of record for a critical process, high switching costs, regulatory or audit requirements and an AI roadmap that already generates revenue. The more of these a company has, the more AI is a tailwind for its valuation.
- How is the score calculated?
- Six weighted factors: product role (22%), AI roadmap (20%), proprietary data (18%), pricing model (16%), switching costs (14%) and regulation (10%). Each answer carries an exposure score; the weighted total runs from 0 (resilient) to 100 (severe exposure).