Model access
Choose which connected apps and teams can use each approved model or provider.
Finn AI Gateway
One place to control how connected applications use AI. Choose the models, set access and budgets, and see usage across Finn and the other AI apps you connect.
AI Gateway
Apply shared rules to AI requests routed through your Gateway. Teams get the models they need. You get visibility into how they are used.
Choose which connected apps and teams can use each approved model or provider.
Track requests and spending by connected application or team. Set budgets and enforce the limits you choose.
Route work to approved models for its needs. Define when an allowed alternative can be used.
Set which approved providers may receive each class of request. Apply those rules to traffic sent through the Gateway.
One AI request, with shared controls
Connect an application to the Gateway to put its AI requests under the same access, routing, and spending rules.
A Finn agent or another connected AI application asks to use an approved model.
The Gateway applies the application’s access, data rules, and budget limits.
Eligible work is routed to an approved model. Requests outside the rules are blocked.
The result returns to the application and usage is recorded against the connected app and team.
The Gateway extends beyond Finn’s agents. Connect internal tools and other AI applications to the same controls. It covers requests routed through it; connecting Finn does not automatically bring unrelated AI traffic under its control.
Common questions
Your business chooses approved models and providers, sets access for each app and team, and applies budgets and data rules to requests sent through the Gateway.
The Gateway covers requests routed through it. Connect each Finn or other AI application that should use its access, data, and spending rules.
Business actions have separate permissions and approval requirements. Access to an AI model does not by itself authorize an agent to change a record or complete a transaction.
Contact the Finn team to discuss hosting, data location, and model processing for your selected setup. Ask for the applicable terms on model training, retention, and deletion as part of that discussion.
Meet Finn
Bring one workflow. Make room for what comes next.