Microsoft Copilot Studio Pricing 2026
$200/tenant/month for 25,000 messages. Copilot Studio lets businesses build custom AI agents in Microsoft 365. Here's the full pricing model, message consumption explained, and when it's worth it.
What Is Copilot Studio?
Microsoft Copilot Studio is the platform for building custom AI agents in the Microsoft 365 ecosystem. It was previously called Power Virtual Agents and was relaunched under the Copilot brand in 2023.
Unlike M365 Copilot (which helps individual users in Word, Outlook, Teams), Copilot Studio lets IT teams and developers build purpose-built AI chatbots and automated agents deployed organisation-wide.
Common use cases
- IT helpdesk bot (password resets, access requests)
- HR FAQ assistant (policies, benefits, onboarding)
- Customer service chatbot on company website
- Sales support bot connected to CRM data
- Internal knowledge base search agent
- Automated workflows triggered by natural language
Copilot Studio Pricing Model
Includes 25,000 messages per month
Additional message packs purchased on top
Some Copilot Studio capacity included for $30/user/mo licence holders
| Monthly messages | Cost breakdown | Total monthly cost | Cost per message |
|---|---|---|---|
| 10,000 msgs | Well within base plan | $200 | $0.0200/msg |
| 25,000 msgs | Exactly base plan limit | $200 | $0.0080/msg |
| 35,000 msgs | Base + 1 pack (10k) | $350 | $0.0100/msg |
| 50,000 msgs | Base + 2.5 packs | $575 | $0.0115/msg |
| 100,000 msgs | Base + 7.5 packs | $1,325 | $0.0132/msg |
| 250,000 msgs | Base + 22.5 packs | $3,575 | $0.0143/msg |
What Counts as a "Message"?
1 message = 1 conversational turn
- One user question + one bot response = 1 message
- Simple FAQ answers (stored responses, no AI) = 1 message
- Single-step API lookups = 1 message
- Basic greetings and menu navigation = 1 message each
More complex operations = more messages
- Multi-step agentic flows count a message per action
- Generative AI responses (AI Builder) consume additional messages
- Chained flows that call multiple APIs count each call
- Complex copilot actions can consume 5–20 messages per user interaction
Who Should Consider Copilot Studio?
- Customer service teams wanting custom support bots deployed in Teams or web
- HR departments handling high-volume repetitive employee FAQ queries
- IT teams automating Level 1 helpdesk (password resets, access requests)
- Orgs already in Microsoft 365 wanting AI automation without custom dev
- Teams of 50+ where message volume justifies $200/mo base cost
- Organisations under 50 employees — the per-message cost is high relative to scale
- Teams that need high-volume automation (per-message pricing scales poorly)
- Businesses not using Microsoft 365 (Copilot Studio is deeply M365-integrated)
- Simple chatbot needs — cheaper alternatives exist (Intercom, Freshdesk, etc.)
- Developers who want full code control — Azure OpenAI Service gives more flexibility
Copilot Studio vs Building Directly on Azure OpenAI
| Factor | Copilot Studio | Azure OpenAI Direct Build |
|---|---|---|
| Setup complexity | Low — low-code, drag and drop | High — requires developer expertise |
| Time to first working bot | Hours to days | Weeks to months |
| Cost at low volume (<25k msg/mo) | $200/month flat | Lower (pay per token only) |
| Cost at high volume (>250k msg/mo) | $3,575+/month (per-message) | Lower with reserved capacity |
| M365 integration | Native — Teams, SharePoint, etc. | Custom connectors required |
| Customisation ceiling | Limited by platform | Unlimited |
| Ongoing maintenance | Managed by Microsoft | Your team's responsibility |
Recommendation: For most mid-market Microsoft 365 organisations building internal automation bots, Copilot Studio is the right tool — the setup speed and M365 integration justify the per-message cost at moderate volumes. For high-volume customer-facing deployments or organisations with strong engineering teams, building on Azure OpenAI directly gives better economics and more control at scale.