Microsoft opens its Copilot harness to developers via Foundry

By Infinithink Studio · October 29, 2026

Microsoft announced on 28 October 2026 that its Copilot harness, used across its own products, is available through Microsoft Foundry. The post targets developers and enterprise customers building agents on several models.

What happened

On 28 October 2026, Microsoft published a post on its official blog introducing the Microsoft Copilot harness. According to Microsoft, the harness is the layer that sits around an AI model and gives an agent the instructions, context, execution environment and controls it needs to complete work. Microsoft says the model supplies the intelligence, while the harness is everything around it.

Microsoft states that this harness was built by the GitHub engineering team and has been adopted company-wide. According to the post, it powers Copilot Code, Copilot Studio, the GitHub Copilot app, the command line interface, VS Code and other products. As of the post, Microsoft says it is making the harness available through Microsoft Foundry, as a fully managed way for developers and enterprise customers to build agents that reason, use tools and complete long-running tasks.

Microsoft also describes some technical details. It says the harness works across model families, with more than 20 frontier models plus the option to bring your own key, and that it can route work to the right model based on the task. The company says it recently ported the harness runtime to Rust, which it credits with reducing session overhead and time to first token. It also says the harness supports open standards such as MCP and agent skills, as well as agent plugins.

On the way it is delivered, Microsoft says the SDK can be downloaded or the harness can be run in the cloud, managed by Foundry. In that setup Microsoft runs the agent loop and the sandbox, while the customer brings instructions, skills, tools and the choice of model. The company adds that its platforms remain open and interoperable, so other harnesses and frameworks can still be used.

Why it matters

The announcement points to a theme that Microsoft itself puts at the centre of the post. According to Microsoft, the AI space moves fast, a leading model today might fall behind tomorrow, and building and maintaining your own harness or model router is slow and hard. The company's answer is to let customers build on the same harness it uses in its own products.

It is worth separating what is announced from what is claimed. The availability through Foundry and the list of products that use the harness come from Microsoft's own post. Statements about performance, satisfaction among developers and battle-tested reliability are Microsoft's own descriptions, and the post does not include figures that readers can check independently.

For anyone following the market, the practical signal is the idea of working with several models rather than one. Microsoft frames the ability to choose the right model for each job as a core advantage of its approach.

What it means for your business

This announcement is aimed at developers and enterprise customers who build agents on Microsoft Foundry. Most small businesses, professionals and agencies will not build an agent harness themselves, so it is reasonable to expect that the effect on them will be indirect, through the tools they already use.

What this suggests is a broader pattern. Tools that let you pick the model that fits each task, rather than tying every task to a single model, are becoming a normal part of how AI products are described. For a team, the useful question is less about which harness sits underneath and more about which tasks are worth handing to an assistant, and who reviews the result.

It is also reasonable to expect that agents will be described more often as systems that follow instructions, use tools and work on longer tasks. A business does not need to adopt that vocabulary to benefit from it. Clear instructions, a defined tone of voice and a person who checks the output matter whichever platform is used.

Finally, since the details above are Microsoft's own, any decision about adopting the Foundry offering is best made after reading Microsoft's documentation and checking the terms directly with the provider.

What to do now

• Read the original announcement on the Microsoft blog, linked in the sources, and note which parts concern developers and which concern end users.
• List the three recurring tasks that take your team the most time, for example replying to emails, drafting posts or summarising documents.
• For each task, write the instructions you would give a colleague, including tone, length and what must never be left out.
• Decide who reviews the output before it goes to a client. The assistant prepares a first draft, the person decides.

If you want to try this on your own material, you can start free with the writing tools on Infinithink, such as the email reply tool or the text summary tool, and keep the final review in your hands.

Sources

• Microsoft, Introducing the Microsoft Copilot harness: Unlocking the potential of truly multi-model agentic work
• The Verge, Microsoft is ending Copilot support on smart TVs

For information only, this is not legal or professional advice. Information is current as of the publication date and may change. Official sources checked on October 29, 2026.