AI Studio
The platform's AI — agents, grounded answers, insights, and automations over your tenant's real data, growing in autonomy across four stages.
AI Studio is the platform's AI. It works on your tenant's real commerce data — products, orders, customers, pricing — rather than a chatbot bolted on the side. It lives in Cockpit, and it grows in autonomy across four stages, so you decide how far to let it go.
AI Studio is rolling out in Cockpit now. The programmatic API — for building on its agents and tools directly from your own code — is coming. This page describes what AI Studio does and what you'll be able to build against; it will gain endpoints and SDK usage when the API ships.
The four stages of autonomy
Each stage builds on the one before it. You start with answers and move toward action at your own pace.
- Assistant — answers grounded questions about your data, with sources. It looks things up, checks availability, and suggests the next step.
- CoPilot — surfaces opportunities before you ask: cross-sell, pricing that has drifted out of its corridor, accounts at risk.
- AutoPilot — acts within the rules you define, with approval where you want it: pricing adjustments, campaigns, reactivation.
- Agent teams — multiple agents coordinating toward one goal across segments, clusters, and channels.
What's inside
- Agents and conversations — assistants you configure for a task, with the conversation history that goes with them.
- Knowledge — your own data and documents, indexed so answers are grounded and cite their sources.
- Tools — the actions an agent can take. Agents call tools to read and act on platform data; you can extend the set with your own.
- Insights — the findings CoPilot raises about your business, with the data behind them.
- AutoPilot rules — the boundaries you set for what AI may do on its own, and where it must ask for approval first.
- Traces — a record of what each agent did and why, for review and audit.
- Credits — usage is metered in credits, so spend is visible and predictable.
For developers
Today AI Studio is operated through Cockpit. When the API ships, you'll be able to build on the same capabilities from your own code — for example:
- Call an agent from an App or an Integration Studio workflow to enrich, classify, or summarize data in line.
- Register your own tools so agents can act through your App's capabilities.
- Surface AI insights inside your own Storefront or admin experience.
Until then, watch the API Explorer and the Changelog — AI Studio endpoints will appear there as they become available.