Photo: Microsoft.
Microsoft has announced plans to allow businesses to create their own AI autonomous agents within Copilot Studio.
Autonomous agents are custom-built AI apps that can handle specific tasks for enterprise workers without supervision, freeing up the employees’ time to handle other projects.
The case for making your own AI Agent
Creating an agent isn't hard. Photo: Microsoft.
According to Microsoft, companies like McKinsey & Company, Pets at Home, and Thomson Reuters are already creating autonomous agents to increase revenue, reduce costs and scale impact:
- British pet care business Pets at Home created an agent for its profit protection team to more efficiently compile cases for skilled human review, which could have the potential to drive a seven-figure annual savings.
- McKinsey & Company is creating an agent that will speed up the client onboarding process. The pilot showed lead time could be reduced by 90% and administrative work reduced by 30%. The agent automates complex processes, such as identifying the right expert capabilities and staffing teams, and acts as a single place where colleagues can ask questions and request follow-ups. Streamlining tasks and reducing manual inputs, this agent could potentially save consultants many hours, allowing them to spend more time with clients.
- Thomson Reuters built a professional-grade agent to speed up the legal due diligence workflow, with initial testing showing some tasks could be done in half the time. This complex workflow requires significant legal expertise and specialist content, and the ambition was to make this process more efficient. The agentic solution combines the knowledge, skills, and advanced reasoning of Thomson Reuters CoCounsel, and integrates into Microsoft’s productivity tools to take on key tasks in this workflow. By significantly transforming the way work happens, the solution helped Thomson Reuters lawyers close deals more quickly and efficiently for their clients.
Photo: Microsoft.
The best part about these autonomous agents is that they can be built with Microsoft’s Copilot Studio using low-code or no-code instructions, allowing users to create AI-driven workflows without requiring extensive technical expertise.
It starts by giving the agent a purpose like telling it to sort and process incoming emails. From there you can specify how the agent should respond it data is missing, or if information is incomplete. Microsoft is also allowing employees to monitor and check the agent’s work for accuracy and deal with any errors, also called "hallucinations".
How you monitor your agent's performance. Photo: Microsoft.
Agents can automatically respond to signals across your business and initiate tasks. They can be configured to react to events or triggers without human input that instead originate from various tools, systems, and databases, or are even scheduled to run hourly, daily, weekly, or monthly.
Once these agents are created, IT administrators can apply a comprehensive set of features to govern and manage their use. The organisation itself is equipped with enterprise data protection including a comprehensive set of features to secure agents, including encryption, data loss prevention measures, and a responsible AI strategy to mitigate AI-related risks such as prompt injection.
Agents built in Copilot Studio will have guardrails and controls established by maker-defined instructions, knowledge and actions. The data sources linked to the agent adhere to stringent security measures and controls that are all managed in Copilot Studio. These include data loss prevention, robust authentication protocols, and more.
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.