Migration Path to Microsoft Copilot Studio

A documented, supported off-ramp for organizations that outgrow Outermind and standardize on Microsoft's enterprise agent stack (Microsoft Copilot Studio and the broader Agent 365 platform).

Why this page exists

Outermind commits to zero lock-in. We are the SMB on-ramp for agentic AI on Microsoft 365: you can get a working agent live in about 30 minutes, without an Azure subscription, a Power Platform administrator, or a six-figure annual commitment. Some customers grow into a dedicated IT function, a Microsoft Partner relationship, and requirements (500+ seats, sovereign cloud, Purview-grade governance) that are best served by Microsoft's first-party stack. When that day comes, we want your graduation to be smooth rather than a rebuild from scratch.

This page documents how the concepts you configure in Outermind map to their equivalents in Copilot Studio and Power Platform, what Outermind exports and in what form, and what you will need on the Microsoft side. It is a documented mapping plus exportable configuration and data. It is not a one-click, automated importer: Copilot Studio agents, Power Automate flows, and Dataverse or SharePoint knowledge sources are built inside Microsoft's tools. Outermind's job is to hand you an accurate inventory and a clear translation so you (or your Microsoft Partner) can rebuild quickly and confidently.

The short version: Everything you build with Outermind is designed to translate. Your agent personas map to Copilot Studio agent definitions, your routing rules to Power Automate triggers, your Grounded Knowledge Indexes to knowledge sources, and your Custom API Tools to custom connectors. We hand you the export and the mapping when you are ready.

Concept mapping

The table below maps each Outermind building block to its closest Copilot Studio or Power Platform equivalent. Use it as the blueprint for your rebuild.

Outermind concept	Copilot Studio / Power Platform equivalent	Notes on the translation
Agent personas and instructions (Personal Assistant, AI COS, SME, Standard Agent; their system instructions and behavior)	Copilot Studio agent definitions (agent name, description, and instructions)	Personas translate cleanly as agent definitions. Your written instructions carry over as the agent's instructions; you re-author them in the Copilot Studio agent editor rather than importing a binary.
Routing rules and triggers (which agent handles which mailbox, schedule, or event; wake-up timers)	Power Automate flows and Copilot Studio triggers	Event and schedule routing becomes Power Automate flows or agent triggers. The logic is documented in your export; the flows themselves are authored in Power Automate.
Grounded Knowledge Indexes (Azure AI Search-backed retrieval over your documents and sites)	Copilot Studio knowledge sources	Because Outermind already grounds on Azure AI Search, the underlying pattern is familiar. In Copilot Studio you re-register the same sources (SharePoint, files, or a connected Azure AI Search index) as knowledge sources and re-index them in the target environment.
Custom API Tools (external HTTP endpoints registered through the wizard, with their auth and JSON schema)	Copilot Studio custom connectors and actions	Each Custom API Tool corresponds to a custom connector or agent action. The endpoint, method, authentication mode, and parameter schema are exported as structured records so your Power Platform admin can define the matching connector.
Escalation and approval flows (Escalation Router, Safety Gateway review queue, Agent Self-Service approvals)	Copilot Studio approval patterns (Power Automate approvals, human-in-the-loop steps)	Your escalation targets and approval gates map to Power Automate approval actions and human-in-the-loop steps in the agent flow. The routing and thresholds are documented; the approval steps are rebuilt in the flow designer.
Model and work-type assignments (which model handles reasoning, composition, safety, and so on)	Copilot Studio model selection and generative orchestration settings	Copilot Studio manages model choice within Microsoft's own model catalog, so this is a re-selection rather than a like-for-like transfer. Your Outermind work-type assignments give your admin a clear starting point.
Audit history (agent execution logs, tool-call records, cost history)	Dataverse tables, Purview audit, and Copilot Studio analytics	Your historical audit data exports as structured records for your own retention and analysis. Going forward, Copilot Studio's activity flows through Microsoft's own analytics and Purview audit.

What Outermind exports, and in what form

When you request a migration export, Outermind provides two kinds of output. Being precise about which is which saves your team time.

Structured, machine-readable exports (files you keep):

• Agent configuration. For each agent: name, type, persona, full instructions, and its model and work-type assignments, exported as a structured configuration file. Related capability is described under Agent configuration and specs.
• Knowledge source inventory. A manifest of every Grounded Knowledge Index: its sources (SharePoint sites, document libraries, uploaded files), the retrievers, and the indexing scope. See Grounded Knowledge Indexes.
• Tool and connector definitions. For each Custom API Tool: endpoint URL, HTTP method, authentication mode, and the request and response JSON schema, exported as structured records.
• Audit and execution history. Your agent execution logs, tool-call records, and cost history, exported as structured data for your own retention.

Documentation you re-implement in Microsoft's tools:

• The Copilot Studio agents themselves. The exported configuration is the blueprint; the agent is created in the Copilot Studio agent editor.
• Power Automate flows. Your routing rules, triggers, and schedules are documented so your team can author the equivalent flows.
• Knowledge source re-indexing. Sources are re-registered and re-indexed in the target Power Platform environment; content is not copied into Dataverse by Outermind.
• Custom connectors and approval steps. Built in Power Platform from the exported definitions.

We deliberately do not claim an automated, one-click migration. Copilot Studio and Power Platform are configured inside Microsoft's own tooling, and a responsible cutover is a rebuild guided by an accurate inventory, not a silent import. What Outermind guarantees is that the inventory is complete and the mapping is honest, so the rebuild is measured in days rather than a from-memory reconstruction.

Prerequisites on the Microsoft side

Before you begin the rebuild, your organization needs the Microsoft foundation that Copilot Studio assumes:

• Copilot Studio licensing. A Copilot Studio license or the pay-as-you-go / message-pack capacity model, in addition to your Microsoft 365 licensing. Copilot Studio is a separate product from the Microsoft 365 Copilot chat assistant.
• A Power Platform environment. A Power Platform environment (with Dataverse) in the region where your agents will run, provisioned and governed by your administrator.
• A dedicated administrator. A Microsoft 365 or Power Platform administrator who can create environments, register custom connectors, grant connector consent, and manage Dataverse. Copilot Studio's configuration surface assumes this role exists.
• An Azure subscription (as needed). If you are connecting an existing Azure AI Search index or other Azure resources as knowledge sources, you will need the corresponding Azure subscription and access.

If your organization has already completed the Agent 365 enablement ceremony with Outermind, your administrators are familiar with the Entra consent and Exchange authorization patterns that Copilot Studio also relies on, which shortens the ramp.

When Copilot Studio is the right destination

Outermind positions itself as the on-ramp, and we say so plainly: Copilot Studio is often the right destination for a maturing organization. Consider migrating when several of the following are true:

• Scale. You are past roughly 500 seats, and per-seat Microsoft licensing plus consumption-based agent runtime is already part of your budget.
• A dedicated Microsoft administrator. You have someone whose role is to run Power Platform, manage environments, and own connector governance.
• Power Platform competency. Your team already builds and maintains Power Automate flows, custom connectors, and Dataverse solutions, so the agent stack is an extension of existing muscle rather than a new discipline.
• Sovereign or government requirements. You need GCC, GCC High, DoD, or a sovereign cloud, or Purview-grade DLP and Defender threat detection on agent activity as a day-one requirement. These are Microsoft first-party capabilities.

If instead you are a smaller team without dedicated IT, you are usually better served staying on Outermind: the flat pricing, the wizard-driven surface, and the 30-minute time-to-value are shaped for exactly that stage. There is no penalty for staying, and no penalty for leaving. When you graduate, this page and your export are how we help you land.

Requesting your export

To start a migration, contact Outermind support from the Help & Support center and request a Copilot Studio migration export. We will generate the structured configuration, the knowledge source inventory, the tool and connector definitions, and your audit history, and we will walk your team (or your Microsoft Partner) through the concept mapping above.

Frequently asked questions

Does Outermind delete my data when I migrate?

No. Your configuration and history remain available in Outermind until you choose to close your account. Exporting for migration does not disable your agents or remove anything.

Can Outermind rebuild the agents inside Copilot Studio for me?

The agents are created in Microsoft's tooling, typically by your administrator or your Microsoft Partner. Outermind provides the export, the mapping, and guidance; we do not have write access to your Power Platform environment.

Will my instructions and personas work identically in Copilot Studio?

The instructions translate directly and are a strong starting point. Because Copilot Studio uses Microsoft's own model catalog and orchestration, expect to tune behavior and model selection after the initial rebuild, just as you would when moving any agent between platforms.

What happens to my Custom API Tools that use bespoke authentication?

Each tool's endpoint, method, auth mode, and schema are exported as structured records. Your Power Platform admin registers the matching custom connector and supplies the credentials or connection references in the target environment.

Is there a fee to export?

No. The migration export is part of our zero-lock-in commitment.