As artificial intelligence (AI) goes from a tool to a real assistant, his role in the workplace develops, fundamentally transforming the operation of companies. Microsoft’s latest research identifies a new type of organization known as Frontier Firm, where intelligence requirements are managed by hybrid teams of AI agents and humans.
CNET: IA Essentials: 27 ways to operate AI for you, according to our experts
On Wednesday, Microsoft published its annual report on the working trends index in 2025, which combines survey data of 31,000 workers in 31 countries, Microsoft 365 productivity signals, LinkedIn hiring and work trends and expert ideas to give employees and business leaders a full view of the work landscape.
This year’s report is the concept of a border business, which helps researchers explore what the work structure in the AI era will look like. As expected, today’s differences are large. However, the report found that each organization will be along the way to become a border business over the next two to five years.
Also: The best AI for coding in 2025 (and what not to use)
“We went through this exercise as a team, where we were like, if we were starting today, it was the first day of the history of work, there was no history that preceded it, how could we conceive of work?” said Alexia Cambon, researcher in the future of work at Microsoft. “We were surprised by the number of hypotheses we had about how the work was to be done.”
The journey
The report identifies three phases for the transformation of an organization into a border business. The first phase implies that employees interact with assistants to work more effectively. In phase 2, humans use AI agents, which work as more a “digital colleague”, performing tasks under human direction and releasing their time.
“The phase in which most of us are at the moment is that AI appears as an assistant at work,” said Cambron. “But we certainly see that the trip will affect the border company, where agents are part of the workforce and help you do new things, new types of net knowledge, and these are two very different types of organizations.”
Also: Why the new OpenAi AI agent tools could change the way you code
The final phase (phase 3) involves Humans working with a team of AI agentswhich can perform whole trade processes and workflows. According to the report, this evolution of AI in knowledge work will imitate how the development of software compatible with AI has progressed general aid in coding to discuss interfaces and on agents that can do the task in the name of the user.
Productivity gains
The report has identified real productivity gains in the implementation of AI In organizations, with one of the largest capacity gap – up to 80% of global workforce, employees and managers, say they have too much work to do, but not enough time or energy to do so.
Now, organizations can “buy information on TAP”, which is made possible through AI agents who act as digital work and help companies to evolve if necessary, depending on the report.
The report revealed that 46% of managers say that their companies use agents to automate workflows or processes. The exact details of the interaction of the human agent will be different by function, certain tasks managed independently by agents and others requiring more human involvement.
However, the addition of AI is not enough, because speed companies evolve exceed the rate at which people work, with a greater change to rethink knowledge based on knowledge.
Also: AI agents are not only assistants: how they change the future of work today
According to the report, business leaders must separate workers from knowledge of knowledge, recognizing that humans who can perform higher level tasks, such as creativity and judgment, should not be trapped by responding to emails. On the contrary, in the same way that the working professionals say that they send emails or create pivotal tables, they will soon be able to say that they create and manage agents – and border companies display the potential possibilities of this approach.
The report revealed that only 844 employees on the sample of 31,000 people worked in companies that met the five traits that include border companies: the deployment of AI on the organizational scale, the advanced maturity of AI, the current use of agents, the use of projected agents and the conviction that agents are essential to achieve the return on investment on AI.
Among these companies, 71% of workers said their business was booming, compared to only 37% worldwide; 55% say they can take more work compared to 20% worldwide; And 90% point out that they can do more significant work compared to 73% worldwide. These workers are also less likely, 21% against 38% worldwide, to fear that AI will take their work.
Labor changes
Beyond the positive results of the AI, there will be inevitable changes in the appearance and functioning of the workforce. Each industry will know the transformation differently, with a defined AI to generate new jobs and replace the others.
Also: As AI agents are increasing, it becomes the new HR department
Among the business leaders questioned, 45% said that the expansion of the team’s capacity with digital work is an absolute priority over the next 12 to 18 months. The three main areas of accelerated IA investment include customer service, marketing and product development.
Even after expanding their digital workforce, researchers found that key component companies must consider to maximize the advantages include Get the right human agent ratio. This concept refers to the removal of a balance between having too few agents per person and undermining their potential, or crushing human capacity with too many agents per person.
As an example of balance, the report quotes a Harvard study which found that an individual with AI surpasses a team without technology, but a team with AI surpasses everyone. Cambron said that this trend highlighted how an organization would get better results when it has a larger team associated with AI, rather than using technology to completely reduce humans.
“I don’t think we have to reduce the workforce and replace it with AI. I think we have to maintain the workforce and increase it by IA to get the best results,” Cambron said.
Another element of the change of work is to see AI as a teammate rather than a tool, allowing technology to take higher level tasks such as Project management. However, 52% of respondents currently consider AI as a tool based on orders and 46% consider it a thinking partner.
“We must work to have a more uniform understanding of how to use it, because if we do it as a search engine or a spreadsheet, it will not make the same return as if we start to use it more as a digital colleague, where we are and we are thinking together,” said Cambron.
The maximization of the performance of AI agents is to learn management problems, including delegation, iteration, demand and refining of technology. The need for someone to manage these teams of AI agents has led to the evolution of a new role, the boss of the agent, responsible for the best performance.
“Everyone will have to manage agents,” said Cambron. “I think it’s exciting for me to think that you know, with agents, each person at the start of their career will be able to experience the management of the first day, from their first job.”
Also: Employers want workers with AI skills, but what exactly does that mean?
Consequently, this change of functions, the traditional organizational table can also see certain changes, replaced by a work table, which Microsoft describes as a “model focused on the results where the teams are formed around objectives, and not functions, fed by agents who extend the scope of employees and allow to work more quickly and more impact to work”.
Microsoft compares this work system to films production sets in which tailor -made teams meet to carry out the project and dissolve once the work is finished. In the same way, various AI agents will work together to achieve a objective quickly, then go to a new project without having to reorganize real employees.
Humans with AI skills are also in demand, with 47% of business leaders listing their existing workforce as an absolute priority over the next 12 to 18 months, 51% of managers claiming that the training or implementation of AI will become a key responsibility for their teams in the five years, and 35% by planning to hire AI trainers to guide the adoption of employees Next 12 months.
Although 33% of business leaders questioned envisage a reduction in staff, Microsoft predicts that new roles will evolve. Most business leaders, 78% for non -frank and 95% companies for border companies, plan to hire people for AI -specific roles, such as IA agents and AI strategists in marketing, finance, customer support and consultation, to prepare for the future.
“We really see this new type of intelligence as additional substitutions, not substitutions,” said Cambron. “I believe that some jobs will disappear, it is a natural order of events, but we absolutely think that we are not going to lack work – I think new jobs will be created, and current jobs will evolve.”