Join our daily and weekly newsletters for the latest updates and the exclusive content on AI coverage. Learn more
Microsoft has announced today a major extension of its artificial intelligence tools with the “spring version of the Copilot Wave 2” wave “Microsoft 365”, introducing new AI “agents” designed to operate as digital colleagues that can perform complex work tasks thanks to deep reasoning capacities.
In an exclusive interview, Aparna Chennapragada, product manager of experiences and devices at Microsoft, told Venturebeat that the company is heading for a vision where AI is more than a simple tool – becoming a full collaborator in daily work.
“We have been at the corner of the street for a long time in the AI world,” said Chennapragada. “It started with all the progress of the model, and everyone was really excited about this and the abundance of intelligence. It is now a question of ensuring that intelligence is available for all people, especially at work. ”
The announcement accompanies Microsoft’s work trend index in 2025, a complete research report based on surveys of 31,000 workers in 31 countries, documenting the emergence of what Microsoft calls “border companies” – the restructuring organizations around intelligence fueled by AI and the collaboration of human agent.
How Microsoft’s “researcher” and “analysts” and “analysts” agents provide in -depth reasoning to corporate work
At the center of Microsoft’s vision are two new AI agents named researchers and analysts, powered by the models of in -depth reasoning of Openai. These agents are designed to manage complex research tasks and the analysis of data that previously required specialized human expertise.
“Think about it as you know, as a truly intelligent researcher and a data scientist in your pocket,” said Chennapragada. She described how the researcher’s agent recently helped her prepare for an examination of companies by connecting information between various sources.
“I used it to say, hey, I have an important commercial review to come … Pull all the past meetings, the past emails, determine the data from the CRM, then say:” Give me constructive and net contributions on the way in which I should be able to advance the balloon forward for this meeting, “she said. “Because of the deep reasoning, it made links that I had not thought.”
These agents will be available via a new “agent store”, which will also feature partner agents like Jira, Monday.com and Miro, as well as personalized agents built by organizations themselves.
Beyond the cat: how Copilot becomes the “browser for AI” in Microsoft’s business strategy
Microsoft positions Copilot as a central organizational layer for AI interactions, similar to the way web browsers organize internet content – not just a chatbot interface.
“I consider Copilot as the navigator of the AI world,” said Chennapragada. “On the internet, we had websites, but we had the browser to organize the diaper. For us, Copilot is this organizing layer, this browser for this AI world. ”
This vision extends beyond simple textual interactions. The company introduces Copilot laptops, which allows users to found AI interactions in specific collections of files and meeting notes. A new co -pilot search feature provides business research capacities powered by AI on several applications.
“Today, most of the AI, we have assimilated it to discuss,” noted Chennapragada. “Sometimes I have the impression that we are in the pre-gui back era, where you have this incredible intelligence, and you say to yourself:” Oh, we have a modem of AOL dial-up stuck on it. “”
To combat this limitation, Microsoft provides generation capacities of OPENAI GPT-4O images to business settings with a new creation feature, allowing employees to generate and modify images that comply with the brand.
Professional exhaustion and employee work interruptions: “capacity difference” at the origin of Microsoft IA Focus
Microsoft’s research reveals a significant “difference in capacity” – 53% of managers say that productivity must increase, but 80% of the world’s labor reports have no time or energy to do their job. The company’s telemetry data show that employees face 275 interruptions per day from meetings, emails or messages – an interruption every two minutes during basic working hours.
“There is so much more repression, latent demand for work, productivity and production,” said Chennapragada. “This statistic really stood out for me, that there is so much more repression, latent demand for work, productivity and production. So I see this as an increase, less a job displacement. ”
Research also indicates a change in AI adoption models. Although last year’s adoption was widely directed by employees, this year shows a more downward approach, with 81% of commercial decision -makers saying that they wanted to rethink the strategy and basic operations with AI.
“It is a change between last year, where it was much more ascending and led by employees,” noted Chennapragada. “What this tells us is that there must be a much more descending AI strategy, but also AI products that you deploy in the company with safety, with compliance, with all the railings.”
Rise of the “ boss agent ‘: how Microsoft is considering employees of digital workforce management
Microsoft predicts a fundamental restructuring of organizations around what it calls “work graphics” – more fluid team structures focused on the results supplied by agents that extend the capacities of employees.
This reorganization will require determining the optimal “human-agent relationship” for different functions, a metric that will vary according to the task and the team. The company expects each employee to become an “agent boss” – someone who manages AI agents to amplify their impact.
“For us at Microsoft, it is not enough that 2% of our customers’ company adopts AI, it really brings the whole business. It is then that you get the full productivity gains,” said Chennapragada.
The company’s research shows that managers are currently ahead of employees in the adoption of this state of mind, with 67% of familiar managers with agents, against only 40% of employees.
To help organizations navigate this transition, Microsoft improves its co -pilot control system with new capacities that allow IT administrators to manage organizational agents.
“What’s going on if you have all these (agents) who run? Our customers have asked for it,” said Chennapragada. “What we have built is a co -pilot control system where administrators can watch and say, what is conformity, what is security, what is data confidentiality, what are the agents in the system?” How can I really manage them? “
Research shows that “border companies” The main adoptions on AI outline competitors by large margins
Commercial implications extend beyond productivity gains. Microsoft’s research shows that 71% of workers in “border companies” – organizations at the cutting edge of AI – is in full swing, against only 37% worldwide.
For small and medium -sized enterprises, the democratization of intelligence can level the rules of the game, allowing small teams to operate with capacities formerly reserved for much larger organizations.
While 33% of managers are considering IA -related workforce reductions, 78% also plan to hire for new Roles specific to AI, including AI trainers, data specialists, security specialists and specialists in AI agents.
Linkedin data included in research show that the most important AI startups increased the 20.6% of annual slipper, which is more twice the pace of large technological companies to 10.6%.
“While the holders adapt and the challengers evolve, as we have seen in the boom of the point-point points, the rules of talent and competition are rewritten in real time,” noted the company in its report.
While Microsoft’s new AI tools take place from the end of May, the scene is ready for what Cennapragada calls “the AI navigator”. Just as the previous technological revolutions have fundamentally changed our way of working, the transition to human agent teams promises to transform not only what work is done, but which, or what, does it.
