{"@context":"https://schema.org","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"What can my team actually ask for?","acceptedAnswer":{"@type":"Answer","text":"Anything inside their permissions: a metric across the year, every case flagged above a threshold last week, what happened on a disputed charge, or a full picture of an account before a call. Nobody needs to know which collection holds the data or which Forest Action to trigger."}},{"@type":"Question","name":"Can an AI agent do more than read and write data?","acceptedAnswer":{"@type":"Answer","text":"Yes. The Forest Actions your engineers coded are exposed through the MCP server, so an AI agent can trigger a refund, change a status or run any operation your team built. Each one carries the same permissions and approval gates as a person running it in the Forest UI."}},{"@type":"Question","name":"How do we give an AI agent its own access?","acceptedAnswer":{"@type":"Answer","text":"Create a service account for it in Forest, with the role and teams it needs, and generate an application token. The AI agent connects with that token, sees only what its role allows, and everything it does is logged under its own name."}},{"@type":"Question","name":"Do we need a developer to build an AI agent?","acceptedAnswer":{"@type":"Answer","text":"No. Your ops or product team describes what the AI agent should do in plain language, and the LLM client turns that description into skills that call the Forest MCP server. Your engineers stay involved on architecture and thresholds, not on writing each AI agent."}},{"@type":"Question","name":"Can an AI agent act without anyone approving it?","acceptedAnswer":{"@type":"Answer","text":"Only below the thresholds you set. Actions carry approval conditions and named approvers in Forest, so above a threshold the AI agent proposes and a person signs. The Forest back-end enforces the gate; a prompt does not."}},{"@type":"Question","name":"Isn't this just an API with a chat window on top?","acceptedAnswer":{"@type":"Answer","text":"A raw API gives access. Forest gives governed access on your system of record: every call lands on the same permission model, the same approval gates and the same audit trail as the Forest UI. Building an interface rather than chatting? That is the Gateway API."}},{"@type":"Question","name":"So we never open the Forest UI again?","acceptedAnswer":{"@type":"Answer","text":"Plenty of teams still do, for case work at volume and for anything that reads better on a screen. What moves to the LLM is the ad-hoc half: the question at standup, the spot-check, the pull before a call. Both surfaces run on the same permissions and the same audit trail."}},{"@type":"Question","name":"How do you ensure customer data never leaves our environment?","acceptedAnswer":{"@type":"Answer","text":"The Forest back-end runs inside your infrastructure, alongside your databases. Customer data travels to a provider only at the point where its service is called, and the response is logged back to the case. Data stays in your systems of record."}}]}