Microsoft’s AI bots will take over the tedious office tasks of workers

Microsoft has developed “virtual employees”, capable of performing increasingly complex tasks in the office. This is the latest step in applying artificial intelligence in the workplace.

The company says that the “autonomous agent” bots are sophisticated and can “understand your work and provide support across roles, teams and functions”, and they “can act on your behalf”. The bots can perform daily tasks, such as responding to emails, and can also monitor them. They are able to work around the clock, and can be contacted 24 hours per day.

Satya Nadda, Microsoft’s third CEO in 50 years, announced at an event held in central London that advances in artificial intelligence are causing significant changes in the service industry.

Nadella stated that “for 70 years, the history of computing has been about digitalising people, places, and things.” “We have now a new reasoning tool to make sense of all this.”

He said that the way people interact and communicate with computers has changed. They now use natural language and it’s multimodal, with image, text, and speech, and more context and memory. Improved data and algorithms are doubling AI model performance every six months.

Other companies that are racing to create these types of autonomous agents are Google, Salesforce, Sierra founded by Bret Tayler, chairman of OpenAI and Harvey which focuses on law agents.

Jared Spataro is the chief marketing officer at Microsoft for AI. He compared the agents with robots performing manual labor in factories. He said that he believed that each organisation would have a mixture of humans and robots doing important work. “But increasingly, these agents will work under the supervision and reporting to humans of humans.

McKinsey was one of the first companies to test this technology. It has created an agent which can help their teams process new proposals faster. It uses company data to create and send email responses.

Now, companies can create “agents”, which are tailored to their business. The first ten agents are available for various jobs, such as case management or sales. Microsoft says that they can “scale up your team’s capacity like never before”.

The company claims that when given instructions and the ability to access all the information of the company, an agent can “reason”, and grasp the scope of the task, to determine what is needed. If the agent gets stuck in complex situations, it will “escalate”, or escalate, a question to an actual human.

Microsoft Copilot, an AI tool, is already used by 2.1 million people each month. It provides services like transcription and summaries for video calls, emails, and document drafting. The new product, which was launched earlier this year, requires far less guidance.

Ece Kamar is the managing director for AI Frontiers, Microsoft. He described the relationship between autonomous robots and their human co-workers as “a collaborative effort”. “My goal personally is not that these agents take over jobs but to create the right interactions and workflows so the agents can be able to handle the boring tasks people don’t like to do and humans can focus on the more interesting aspects of their work.”

She said that in the future “it will be very clear to see how these actions could have negative consequences”. We will also need to examine and investigate how we can put in place the correct transparency layers, preventions and stopgaps.

Spataro claimed that the agents will transform the workplace. The organisation we know today is really a product of two world wars. You will find the same things in every firm today: an HR department and a sales department. There is also a marketing department and a finance section. “Almost all firms on earth look the same,” said he.

We think that in the future, more generalised leaders are going to be able do more specific jobs. We’ve observed that flatter structures and generalised skills are more common than the specialisation we’ve seen in the past two hundred years.

The agents should be able, through their own work and interaction with humans and each other, to act and take decisions without any supervision.

The launch of Agents is yet another triumph for the tech giant, which was quick to steal the march on its rivals and transform generative AI into an usable product that businesses can use, eager for its productivity gains. Microsoft is a partner of OpenAI and ChatGPT. Microsoft has invested in so-called Frontier Labs that are building large language models like Mistral.

Microsoft hired Mustafa Suleyman in March, along with his Inflection AI team, to create a new organization called Microsoft AI. The organisation is focused on improving Copilot.

These relationships with leading AI companies have attracted the attention of antitrust regulators both on the US and European sides. The Federal Trade Commission in the United States has launched an investigation into Big Tech’s investments and partnerships with generative AI, including the partnership between Microsoft and OpenAI. Microsoft’s deals with Inflection, Mistral and OpenAI will not be subject to further scrutiny by the UK Competition Authority.

Microsoft’s market capitalization has increased by 27 per cent over the last year. In its most recent quarterly earnings, published in June, the company reported revenue at $64.7 billion up 15% and net income at $22 billion up 10%.