Your assumption is correct- the cost for using RAG in Agent Copilot is covered in the $40/60 price paid for each Agent Copilot user.
Original Message:
Sent: 03-03-2026 05:11
From: Hichem Agrebi
Subject: Clarification on Knowledge Fabric AI Experience Token Consumption Model
Thanks Amanda, the reference document isn't on resource center but on Genesys Genesys Cloud Knowledge Fabric - FAQ
So this confirms then that there is no limit to using RAG pipelines through knowledge fabric when using Copilot hence all usage costs are included respectively in the 40token/named and 60token/concurrent users?
LLM token consumption is evident for any AI query and the main ask was around associated experience toekn usage as in any add-on cost.
Thanks for your confirmation
------------------------------
Hichem Agrebi
------------------------------
Original Message:
Sent: 03-03-2026 04:36
From: Amanda Halpin
Subject: Clarification on Knowledge Fabric AI Experience Token Consumption Model
Hi Hichem,
The tokens in this sentence refers to LLM tokens, not AI pricing Tokens.
"Each time Knowledge is used to generate an answer, perform reasoning, or execute an action through an AI-powered tool, the interaction consumes a predefined number of tokens based on complexity and compute usage."
Apologies for the confusion, I'll request this to be made clearer. Can you send me the url that you're referencing?
Thanks,
Amanda
------------------------------
Amanda Halpin
Principal Product Manager, Knowledge @ Genesys
Original Message:
Sent: 03-02-2026 04:24
From: Richard Chandler
Subject: Clarification on Knowledge Fabric AI Experience Token Consumption Model
Hi Hichem,
For an official response you'll need to go to your account/partner manager. However there is no separate line item or AI Token item for RAG etc. I have been running a number of Knowledge Fabric and RAG based interactions with no additional charges except the expected AI Tokens.
------------------------------
Richard Chandler
Connect
Original Message:
Sent: 03-02-2026 04:12
From: Hichem Agrebi
Subject: Clarification on Knowledge Fabric AI Experience Token Consumption Model
Thanks Richard,
Appreciate the clarification regarding LLM tokens vs AI Experience Tokens.
However, I'm specifically referring to the wording in the Knowledge Fabric FAQ, which states:
"Each time Knowledge is used to generate an answer, perform reasoning, or execute an action through an AI-powered tool, the interaction consumes a predefined number of tokens based on complexity and compute usage."
And also:
"Knowledge Fabric can only be used through Virtual Agents or Agent Copilot, and any AI consumption through those touchpoints is billed according to their respective usage rates (e.g., interaction-based pricing or AI Experience Tokens)."
I fully understand that LLM tokens are not the same as Genesys Cloud AI Experience Tokens. Any AI query generates input and output or inference token usage anyway and there's an add-on cost for setting up and using a RAG piepline. Why would the response above mention billing if it was LLM tokens that would be included? May be it's my english, but the nuance is pretty obvious to me.
My question is whether the "predefined number of tokens" referenced above refers to customer-billable AI Experience Tokens, or only to internal LLM tokens that are not metered per customer. It is clear that the setup of the RAG pipeline and document sync with data source is not charged but my question refers to the usage and knwoldge surfacing in copilot/VA.
The wording suggests AI Experience Token consumption, but I would appreciate explicit confirmation from Genesys to avoid any misunderstanding. If Genesys confirms that this is all included in GC2/GC3/GC4 licences and covered by the $40/named or $60/concurrent user price then it's perfect and good news for all customers. If however there's an add-on cost, we need a way to measure and track it and budget it somehow upfront to avoid any surprises.
There's also a mention of a limit of 17 sharepoint folders but nor indication at all about any max number of documents (in addition to a missing hint that only text documents are supported).
Thank you.
------------------------------
Hichem Agrebi
Original Message:
Sent: 03-02-2026 03:29
From: Richard Chandler
Subject: Clarification on Knowledge Fabric AI Experience Token Consumption Model
Hi Hichem,
The Tokens in this case are LLM Tokens not AI Experience Tokens. This is referenced in https://help.genesys.cloud/articles/understand-genesys-agent-copilot-ai-models-and-llm-input/ "Large language model (LLM) tokens are not the same as Genesys Cloud AI Experience tokens."
The LLM Token costs are covered by the feature itself and are not metered per customer or passed on. The only time currently where there is some level of correlation is AI Guide creation where 1 AI Token is used which covers the variable underlying cost.
------------------------------
Richard Chandler
Connect