Revenue Growth Alone Doesn't Prove Adoption — Look for a Usage Metric Too

Salesforce's latest earnings paired two numbers that are more convincing together than either is alone: 240% annual recurring revenue growth for Agentforce, and a 97% quarter-over-quarter increase in the actual volume of agentic work delivered. The financial number alone could, in principle, reflect committed contracts and pricing changes without corresponding real usage. The usage number alone doesn't tell you the business is actually monetizing that usage well. Together, they tell a much more complete story — and that combination is worth specifically looking for whenever a vendor makes an AI adoption claim.

Why Revenue Growth Alone Can Be an Incomplete Signal

A software vendor's ARR for a new AI product category can grow rapidly for reasons that don't necessarily reflect deep, genuine usage: bundling the new capability into existing contract renewals, aggressive early-adopter pricing that inflates apparent scale, or customers purchasing capacity ahead of actual rollout. None of these are dishonest or even unusual in early-stage product categories — but they mean ARR growth alone can outpace genuine operational adoption, at least temporarily.

What a Genuine Usage Metric Adds

A usage-based metric — actual work performed, transactions processed, queries handled — measures something contracts and pricing can't directly manufacture: whether customers are actually running the product against real workloads. When a usage metric grows roughly in proportion to (or faster than) the revenue metric, as in Salesforce's case, it suggests the financial growth reflects genuine expanding usage rather than just larger or more contracts being signed without corresponding operational deployment.

What to Ask Vendors For

• Request both the financial metric (ARR, revenue) and a genuine usage metric (work units processed, active usage frequency, transaction volume) for the specific AI product you're evaluating — not just whichever number the vendor leads with
• Compare the growth rates of both metrics against each other — a usage metric growing meaningfully slower than the revenue metric over multiple quarters is worth investigating further, since it can indicate contracted capacity not yet translating into operational use
• Ask how the usage metric is defined specifically — a vaguely defined 'engagement' number is less useful than a concretely defined unit (a completed workflow, a resolved ticket, a processed transaction)
• Track both metrics over multiple quarters rather than relying on a single snapshot, since a temporary divergence between financial and usage growth in any one quarter is less meaningful than a sustained pattern

The Practical Takeaway

When a vendor reports strong AI product revenue growth, ask for the accompanying usage metric before treating the financial number as proof of genuine adoption. Financial growth paired with proportional or faster usage growth is a meaningfully stronger signal of real product traction than either metric would be reported alone.