Why revenue multiples undervalue AI-enabled businesses (and what to use instead)

If you are valuing a business that has deployed AI agents, using a simple revenue multiple is likely undervaluing it. Here is why, and what the alternative looks like.

The problem with revenue multiples

Revenue multiples work by comparing your business to similar businesses that have sold recently. The formula is simple: your revenue times a multiple derived from comparable transactions.

The issue: comparable transactions from 12-24 months ago reflect businesses running pre-AI operations. If your business has deployed agents that have improved margins by 30%, doubled lead conversion, and cut admin costs by 35% — the comparables do not capture any of that.

You are being valued on the old economics of your industry while operating on new ones.

The discounted cash flow alternative

Business Valuation Resources makes the case that the best method for valuing AI-enabled businesses is the discounted cash flow (DCF) approach. The logic is straightforward:

AI's impact is so pervasive across revenue, cost, and HR functions that no single backward-looking metric captures the full picture. A DCF model values the business on its projected future cash flows — which is exactly where AI improvements show up.

Here is what that looks like in practice:

Revenue projections incorporate AI-driven improvements: faster lead response increasing conversion rates, automated upsell sequences lifting average order value, agent-powered outbound growing pipeline volume.

Cost projections reflect automation: reduced headcount for admin functions, lower cost per customer interaction, automated reporting replacing manual analyst hours.

Margin expansion is modelled explicitly: the gap between AI-enabled margins and industry-average margins is the value creation story.

Risk adjustment accounts for the durability of AI improvements: a16z's research shows that agentic workflows create real switching costs and operational stickiness, which reduces the risk of margin compression.

What this means for sellers

If you are selling a business with AI agents in production, a DCF-based valuation tells a more accurate — and typically higher — story than revenue multiples. But it requires evidence:

• Documented before/after metrics. Response times, conversion rates, cost per function, margin trends. The more data you have, the more defensible the projections.
• Production systems, not pilots. A buyer will discount pilot results. Twelve months of production data is worth ten times more than a 30-day experiment.
• Transferable systems. If the AI improvements are dependent on one person's prompt engineering skills, the value is fragile. If they are built into production agents with documentation and monitoring, the value transfers with the business.

What this means for buyers

If you are acquiring a business that claims AI-driven improvements, interrogate the management forecasts. Specifically:

• Are the AI improvements already in production or are they projected?
• What is the cost to maintain the AI systems (compute, monitoring, updates)?
• How dependent are the improvements on specific people or specific vendors?
• What happens to the margins if the AI systems fail or underperform?

The DCF model is only as good as the assumptions behind it.

The practical takeaway

Whether you are buying or selling, the valuation conversation has shifted. Historical revenue multiples are a starting point, not an answer. The businesses commanding premium valuations in 2026 are the ones that can demonstrate — with data — how AI agents have changed their future cash flows.

Start with a valuation estimate to see where your business sits on industry-standard multiples. Then consider what a DCF analysis would look like with your actual AI-driven improvements modelled in.

Sources

• BVR: How Do You Measure the Value of AI?
• Finro: AI Agents Valuation Multiples 2025
• a16z: How 100 Enterprise CIOs Are Building AI in 2025
• SaaS.group: AI Valuation Multiples 2025