We're testing Agentforce Coworker in a sandbox and are trying to estimate Flex Credit cost if we open it to all users. From a single test user, we see a few hundred thousand prompt tokens per month, but most of it comes from one long multi-turn session.
How are other teams tracking per-user consumption and forecasting rollout cost? Specifically:
- Which reports or objects are you using to see credits consumed?
- Do you have a rule of thumb for average credits per user per month?
- Does the unmetered employee-usage add-on cover Coworker?
Thanks in advance.
#Agentforce #Generative AI
I couldn't find a published average credits-per-user figure for Coworker, so I wouldn't forecast from one long multi-turn sandbox session. Officially, Agentforce Coworker supports two models: seat-based (Agentforce 1 Edition or Agentforce for Sales/Service/Industries) and usage-based (Flex Credits, Einstein Requests, or Data Services Credits). On usage-based, CRM/Slack search draws Standard/Basic/Advanced Prompts against Flex Credits (or Einstein Requests on legacy Foundations), and you review that consumption in Digital Wallet. Seat-based/unmetered employee access does cover Coworker CRM and optional Slack search — but only after you assign the Unmetered User Based AI permission set license and permission set; without that assignment those users still consume credits. Extra Data 360 sources are metered separately (Data 360 Flex / Data Services). I haven't seen a first-party per-user usage object called out beyond Digital Wallet's Coworker credit views, so I'd size the rollout from Digital Wallet in the sandbox rather than invent an average. https://developer.salesforce.com/docs/data/agentforce-coworker/guide/agentforce-coworker-billing-considerations.html
