The company is adding "autonomous agents" that are designed to automate business tasks.
They are like digital assistants; in that they are designed to automate and manage specific tasks and processes.
They can be used to sort emails, respond to customer inquiries, schedule follow-up meetings, monitor supply chains, and analyze large amounts of data.
Below is a video showcasing what autonomous agents are capable of:
Microsoft's announcement highlighted how select companies with early access integrated autonomous agents to streamline their workflow processes.
Fraud is a big problem for many companies. UK-based pet care company Pets at Home used Copilot's autonomous agents to compile cases for its profit protection team.
Kay Birkby, senior fraud manager at Pets at Home, said that using Copilot's autonomous agents feature lets them go through vast amounts of data and easily identify patterns of systematic fraud, such as repeated use of the same photo.
"We're finding those anomalies and patterns of behavior, not just to identify undesirable behavior, shall we say, but also to identify cases where this customer experienced an issue that we need to look into further," Birkby told Microsoft. "Is there an issue with this product, for example, that we need to take internally?"
"I am always aware that if I get my job wrong then I can really impact our genuine customers because effectively I could go too hard and block them from being able to transact with us," she added. "It is really a fine line."
The management consulting firm McKinsey uses an agent to streamline the onboarding process for new clients, while law firm Clifford Chance has leveraged one to speed up legal document reviews.
Thomson Reuters built an agent to improve its legal due diligence workflow. Initial tests revealed that some tasks could be completed in half the usual time.
Addressing productivity and potential challenges
Microsoft CEO Satya Nadella emphasizes that autonomous agents are designed to make monotonous tasks easier for organizations to handle and enhance worker productivity, not replace human labor.
Nadella describes Copilot Studio as a "no-code way" to create agents.
However, whether autonomous agents can match human capabilities in every domain remains to be seen. With AI, there are always concerns over "hallucination" in output. For simple tasks, it can excel and streamline workflow processes, but for complex tasks that require a more nuanced approach, is it good enough? Perhaps not yet, but at the rate AI is advancing, in a few years, who knows?