AI’s Next Job: Moving From Solo Tool to Collaborative ‘Digital Chief of Staff’ | PYMNTS.com

This post was originally published on this site.

The role of generative artificial intelligence (AI) in the workplace is shifting from summarizing or creating content for single users to co-working and collaboration with multiple agents, according to the chief product officer of Microsoft Experiences and Devices.

“In the early phase, we were just adding AI to existing applications,” said Microsoft’s Aparna Chennapragada, referencing tools like Microsoft Copilot, which can summarize meetings, draft emails and help with research.

“The second phase is now about co-working and collaborating with AI,” she said. With improvements in deep reasoning capabilities, AI is moving from basic automation to becoming a “thought partner” that can analyze complex problems and assist leaders in making better decisions, she said.

However, the “more interesting” change will come in the third phase, which will be AI’s role in group collaboration, Chennapragada added.

“So far, everything has been single-player, individuals with AI. The third phase is actually a group of people and a group of agents working,” collaborating and co-existing together, she said.

The human employee will interact with an AI agent that acts as a “digital chief of staff,” which in turn will manage a team of other AI agents that each specialize in different tasks or domains, she said during a panel at last week’s HumanX conference in Las Vegas.

This shift will create what Chennapragada describes as a “cognitive energy surplus.” Instead of employees wasting time on repetitive, low-value tasks, AI will handle the routine while human workers focus on creativity, strategy and decision-making.

In a few weeks, Microsoft is planning to launch an initiative combining deep reasoning capabilities with an enterprise’s knowledge graph as a foundation for custom AI agents, Chennapragada said.

Read more: AI Agent Systems Are Here — Will They Transform B2B?

AI Normalization at Work

The use of AI use is becoming increasingly commonplace and accepted in the workplace, Chennapragada said.

Previously, employees often used AI tools without their employers’ explicit approval. A Microsoft study from a year ago found that over 80% of workers were using AI at work, but only around 50% were comfortable telling their managers.

That is changing rapidly. “There’s a lot more acknowledgment that this is actually useful for employees,” she said. “In response, CIOs are saying: ‘We want both — we want our employees to be able to use it, but there needs to be guardrails’” in audit, governance, compliance and security, she added.

Sharing the stage with Chennapragada, Jerry Dischler, president of cloud applications at Google, said CIOs and CTOs he talked to have said that cost is a factor in not expanding AI usage to more employees. But after Google included more features in its base product, there was a “very large” increase in usage, he said.

“They want everybody to use it,” Dischler said. As long as there are guardrails, “they very much see it as part of the strategic capability of their business to get good at AI.”

Despite growing enthusiasm for AI, they acknowledged that companies still face barriers to adoption. Chennapragada pointed to three key hurdles:

1. Information overload — “There’s just this gush … of AI tools. What do we even do? Which of these tools do we choose?” Organizations are overwhelmed by the sheer volume of AI solutions available.
2. Defining next steps — Using today’s generative AI tools, “people can ask questions and get answers, but what’s the next click after that?” Companies may struggle to figure out what to do next.
3. User interface needs work — “The chat interface is fine and all, but it’s like an AOL dial-up modem interface to a superintelligence,” she said. The user interface with AI is a blank box. It needs to improve.

Dischler added that after a company rolls out AI tools to its employees, “everybody just looks at it and (says), ‘I’m not sure what I do with it.’” That is a problem with the user interface by builders like Google and Microsoft. “We should have user interface patterns that allow people to do their best work with the help of AI every single time.”

Both Microsoft and Google executives agreed that AI is on the verge of becoming far more autonomous. “I think we’re going to be surprised by what AI can do with minimal interaction from users over the next 24 months,” Dischler said.

Chennapragada took the idea a step further, predicting that the AI-driven workplace transformation would change job roles altogether.

“In 12 to 24 months, every employee will be a manager — every worker will be the boss of these AI agents,” she said.