AI models and agent platforms · Agent build and run platforms
Microsoft Copilot Studio
Low-code agent builder for Microsoft 365, Teams, Dynamics and Power Platform, billed on Copilot Credits.
Our pick for agent build and run platforms- Vendor
- Microsoft
- Pricing model
- Copilot Credits, pooled across the tenant: prepaid packs or pay-as-you-go.
- Runs on
- Microsoft cloud, inside your tenant
- Agentic OS layer
- Agents
- Industries
- Any industry
- Last reviewed
- November 2026
Our verdict
The default agent platform for Microsoft shops: closest to where people work, over 1,400 connectors, and governance through the tools IT already runs. Credit pricing rewards good design and punishes sloppy agents, so estimate consumption before you scale.
What it is
Copilot Studio is Microsoft's low-code platform for building agents, formerly Power Virtual Agents. Makers design agents that answer from company knowledge, call connectors and Power Automate flows, and run in Teams, Microsoft 365 Copilot, websites and other channels. Agents can be conversational or autonomous, triggered by events.
Billing moved from messages to Copilot Credits in September 2025. Credits are pooled at tenant level, not per user. Employees with a Microsoft 365 Copilot licence can use internal agents without consuming credits, which changes the economics for organisations that already bought Copilot.
Best for
Organisations on Microsoft 365, Teams and Dynamics that want HR, IT, finance and service agents inside the tools people already use, with Power Platform governance.
Not the right fit
Teams outside the Microsoft ecosystem, and complex multi-agent systems that need full code control, testing and custom orchestration; those suit a framework such as LangGraph.
Key capabilities
Low-code agent design
Topics, generative answers, actions and autonomous triggers, built in a visual designer.
1,400+ connectors and MCP
Power Platform connectors, Power Automate flows and Model Context Protocol servers as agent tools.
Grounding in Microsoft Graph
Answers from SharePoint, Teams, email and files the user is allowed to see.
Many channels
Publish to Teams, Microsoft 365 Copilot, websites, mobile and contact centre channels.
Governance
Environments, data loss prevention policies, admin analytics and agent oversight through Microsoft's admin tools.
Credit estimator
Microsoft's estimator models credit use per agent before you deploy.
Strengths
- Lives where Microsoft 365 users already work
- Huge connector library and Power Automate for actions
- Internal agents are free of credits for Microsoft 365 Copilot users
- Governance through admin and DLP tools IT already knows
Watch-outs
- Credits are pooled tenant-wide; one team can drain the shared pool
- Grounded and reasoning responses consume far more credits than simple ones
- With packs, agents stop when use passes 125% of capacity unless pay-as-you-go is on
- Low-code agents get hard to test and version as they grow
Pricing
Copilot Credits cost $0.01 each pay-as-you-go, billed monthly in arrears through Azure, or $200 per month for a pack of 25,000 credits. A pre-purchase plan gives discounts from 5% up to 20% at large volumes. Microsoft 365 Copilot is $30 per user per month for enterprise and lets licensed employees use internal agents without credits.
Consumption depends on design: Microsoft's own example puts a Graph-grounded answer at 12 credits, and far more when reasoning is added. Industry estimates put five departmental agents at several thousand dollars a month in credits.
Checked November 2026. Vendors change plans often, so confirm current pricing with Microsoft before you budget.
Security and governance
Runs inside your Microsoft tenant with Entra ID identity, environment-level controls, Power Platform data loss prevention policies, audit and admin analytics. Agents act with the permissions of the user or a defined service identity.
How we would roll it out
- Model credits firstUse the credit estimator for each agent and decide which users already have Microsoft 365 Copilot licences.
- Separate environmentsBuild in development, test in a test environment, and promote to production with DLP policies set per environment.
- Start with internal agentsHR, IT and finance help agents for licensed users cost no credits and prove value quickly.
- Assign owners and budgetsGive every agent an owner, a credit budget and a monthly review of consumption and outcomes.
Compare with
Other tools in AI models and agent platforms, closest match first.
Amazon Bedrock AgentCore
Agent build and run platforms
Modular AWS services to run, connect, secure and observe production agents, with any framework or model.
Best for Engineering teams running production agents inside AWS.
LangGraph and LangSmith
Agent build and run platforms
Open-source framework for controllable, stateful agents, with LangSmith for tracing, evaluation and deployment.
Best for Engineering teams that want full control of agent logic and to stay cloud-neutral.
ChatGPT Enterprise
Enterprise AI assistants
The most widely adopted general AI assistant, with enterprise admin, residency and compliance controls.
Best for Organisations that want one assistant everyone already knows how to use.
Claude Enterprise
Enterprise AI assistants
Assistant, coding agent and desktop task agent in one seat, with published seat pricing and usage billed separately.
Best for Teams that work with long documents, write a lot, or want coding and desktop agents in the same seat.
Gemini Enterprise
Enterprise AI assistants
Google's agent platform for every employee: enterprise search, prebuilt agents and a no-code agent builder.
Best for Google Workspace organisations that want search and agents across company data.
Choosing between Microsoft Copilot Studio and the alternatives?
Book a call. We look at your systems, data and governance together and tell you which platform fits, and what to build on it first.