Figma’s November 4, 2026 results say that, as of October 31, 65% of paid customers with more than $10,000 in annual recurring revenue were using its agent weekly. That is a customer-level adoption measure. A team deciding how broadly to roll out an AI tool needs a second view: how many people use it and what they accomplish.
Keep the unit beside the percentage
Imagine a fictional software supplier with 100 customer organizations in a defined group. If 65 have at least one weekly agent user, organization-level adoption is 65%. That result could coexist with very different employee usage patterns. An organization with one weekly user counts once, just as an organization with 100 weekly users does. These numbers are illustrative and do not describe Figma’s customer distribution.
Match the measurement dates
Figma separately reported 16,566 paid customers above $10,000 in ARR as of September 30. The weekly agent figure refers to October 31. Multiplying the September count by the October percentage would combine two different snapshots and would not produce a verified count of organizations using the agent. ARR is an annualized contract measure; it is not a count of employees.
Build a small internal adoption table
For your own pilot, record the number of eligible people, the number who used the tool during the same week, and the task each person tried. For example, if four of 20 eligible designers use an agent, weekly person-level adoption is 20%. Ask whether they completed the intended task and how much review was needed before expanding access.
Keep organization-level reach, individual use and task results in separate columns with the same reporting dates. A single active team can establish that a company is trying a tool; it cannot establish that everyone uses it or that the work improved. No Figma accounts or usage data were tested for this article.