Microsoft is unveiling autonomous artificial intelligence agents, or AI employees, capable of performing tasks such as handling client queries and identifying sales leads. This initiative aims to demonstrate to investors that the AI boom can yield indispensable products.
The American technology company is offering customers the ability to create their own AI agents, in addition to releasing ten pre-configured bots designed for various roles, including supply chain management and customer service, under the Microsoft Copilot brand.
Early adopters of the Copilot AI product, which is set to launch next month, include the prestigious consulting firm McKinsey, which is developing an agent to manage new client inquiries by performing tasks such as scheduling follow-up meetings. Other initial users include the law firm Clifford Chance and the retailer Pets at Home.
Microsoft is promoting AI agents, which operate without human intervention, as an example of the technology’s potential to enhance productivity—a measure of economic efficiency, or the amount of output generated by a worker per hour worked.
Satya Nadella, Microsoft’s chief executive, revealed the AI agents at a company event in London, stating that the tool would reduce “drudgery” and increase productivity by freeing up time for more valuable tasks.
“These tools are fundamentally transforming outsourcing, increasing value and reducing waste,” he remarked.
Nadella described MS Copilot Studio, which does not require coding expertise from its users, as a “no-code way for you to be able to build agents”. Microsoft is powering these agents with several AI models developed in-house and by OpenAI, the developer of ChatGPT.
Microsoft is also developing an AI agent capable of conducting transactions on behalf of users. Mustafa Suleyman, the company’s head of AI, mentioned that he has witnessed “stunning demos” where the agent makes purchases independently, although it has also encountered “car crash moments” during development. Suleyman added that an agent with these capabilities would emerge “in quarters, not years”.
Addressing concerns about AI’s impact on employment, Charles Lamanna, a corporate vice-president at Microsoft, told the Guardian that agents would eliminate the “mundane, monotonous” aspects of jobs.
“I believe it is more of an enabler and an empowerment tool than anything else,” he stated.
Lamanna compared the advent of AI tools such as agents in the modern office environment to the arrival of personal computers several decades ago.
“The personal computer did not appear on every desk initially, but eventually it was on every desk because it brought so much capability and information to the fingertips of every employee,” he explained.
“We believe that AI will follow a similar trajectory. It is currently appearing in a subset of departments and processes, but it is only a matter of time before it is present in all parts of an organisation.”
Andrew Rogoyski, a director at the Institute for People-Centred AI at the University of Surrey, suggested that AI agents could help technology companies generate returns for investors who have heavily backed the technology. In June, Goldman Sachs questioned whether a $1 trillion investment in AI over the next few years would “ever pay off”.
“AI companies have consumed a significant amount of investment money and need to generate some returns,” said Rogoyski. “ Assistive agents are a way of demonstrating everyday benefits, although the revenue they will generate remains uncertain.”
However, he cautioned that while agents have been discussed as a concept for years, “we have yet to deliver an agent that is as capable as a human worker”.
Comments