Does Riff use customer conversation data to train shared AI models?

Question

Riff · Accepted Answer

TL;DR

No. Riff does not use customer conversation data to train shared AI models. Conversation data and knowledge base content improve each organization's own buyer graph, which is isolated to that organization through a single-tenant design.

Does Riff use customer conversations to train shared models?

Riff does not use customer conversations to train any shared AI model. Each customer's data reinforces only that customer's buyer graph.

Conversation data and knowledge base content feed the buyer graph of the organization that supplied them.
No customer's data informs another customer's answers or models.
Buyer conversations often contain positioning, pricing objections, and technical requirements, so this isolation is deliberate.

What is a single-tenant buyer graph?

A single-tenant buyer graph is a dataset and model layer built for one organization and used only by that organization. Riff isolates each customer's graph so that its learning stays within that customer's account.

Each graph is specific to one organization.
Riff does not share graph data across tenants.
Answers served to buyers draw only from the organization's own knowledge