AI agents and HR colleagues in the same team - but who will be the boss?
A new metric will be needed when designing teams for the workplace: the ratio of human and AI agents to colleagues in a project. This is the near future and, for some frontier firms that are already moving towards a next-generation organisational model, the present, according to Microsoft's Work Trend Index.

For the Microsoft Work Trend Index report, 31 000 employees and managers in 31 countries were surveyed. We summarise the main findings of the report, and ask Rita Kamasinszki, Director of Microsoft Hungary's Enterprise Business Unit, about the details and practical implications and experiences.
Human-AI hybrid teams are the future
Key findings:
- AI-powered, human-driven systems will be in place to get work done faster and more efficiently.
- Digital colleagues will be full members of teams.
- Humans will fill the leadership role, but the work - even the entire business process - could be done by AI agents.
- 82% of executives surveyed would add digital "workers" to their capabilities in the next year and a half. In part because 80% say they don't have the energy or time to do everything they need to do.
- Work is interrupted 275 times a working day for some reason, which means that on average something interrupts work every 2 minutes, with 60% of meetings being ad hoc, i.e. unplanned. The proportion of out-of-hours messages has increased, and meetings scheduled after 8pm are more frequent due to people working in different time zones. This often leads to chaotic and fragmented work, which is felt by around half of both employees and managers.
New organisational model: what is a "frontier firm"?
It's a new generation of organisational model based on artificial intelligence (AI)-driven operations under human leadership, says Rita Kamasinszki, head of Microsoft Hungary's enterprise business unit. In other words, it refers to those pioneering, leading-edge companies where AI has already been implemented throughout the organisation and is being used in a mature way, there are AI assistants and agents, and the scope of these will expand because they are seen as key to achieving business goals. These companies are not just using artificial intelligence as a tool, but are looking at it as a scalable, always-on service available 24/7. In this setup, it's a collaboration of hybrid human-AI teams.
844 respondents to the Work Trend Index work for such "frontier" firms, which are among the earliest adopters.
71% say their company is soaring, compared with only 37% of global respondents who see this as the case for their own company
.55% say they are able to take on more work as a result (20% in other cases) and also have more opportunities to actually do meaningful work.
93% are optimistic about future job opportunities (vs. 77%) and are much less worried about AI taking their jobs (21% vs. 38%)
Employees in these companies are taking on new roles, delegating tasks to AI assistants and agents. In other words, AI is deeply embedded in operations, which speeds up the work, Rita Kamasinszki outlined. She said this model is already working in the financial, telecom and of course technology sectors. And we already see less complex solutions (cahtbots, other service activities) in everyday life in sales, customer services, internal or HR or IT areas, making work much more efficient and faster.
New dynamics in mixed human - digital teams
"New skills, capabilities and dynamics are needed in organisations where AI agents are effectively team members, not just software. They can autonomously manage tasks, adapt in a timely manner or initiate damage," he stressed. The most important skills and capabilities that will be needed in the future are to guide the AI (prompts, context, giving appropriate guidance) for effective collaboration. Based on his own experience and practice, he stressed the need to treat MI as a partner, evaluate its suggestions, ask for feedback, refine the results. You don't have to accept the first answer, you can work on it together. That is why critical thinking and the role of verification is essential. You have to be able to manage it properly, to override the results if necessary.
"If we think of agents as colleagues, they need to be led by someone. Colleagues need to become agent bosses. If someone becomes a manager, they will need skills such as delegation: what can be delegated to AI, what requires human intervention, and similar issues. It's important to pay attention to these when training employees."
Are jobs at risk if there are more AI colleagues?
"Human jobs will not disappear, but will be transformed in the future by the use of artificial intelligence. AI will automate many routine tasks, but this does not necessarily mean a net reduction in jobs. The types of jobs will be rearranged," says Rita Kamasinszki.
Survey results show that 33% of executives surveyed are considering downsizing in areas where AI is taking over, while 78% plan to hire for new positions focused on artificial intelligence.
Rita Kamasinszki: \"The right ratio will be the key to success.
In addition, there is currently a capacity shortage. 53% of managers said productivity should be increased, while 80% of workers already feel overworked. But AI agents can fill this gap, with workers delegating work to them, while human colleagues can focus on higher value-added - strategic tasks. Companies at the forefront, where some insight into this is already in place, have a more positive view of the future than those where this process has not yet started.
This change may particularly resonate with the needs of the younger generation, because AI will be taking on tasks that were previously left to junior colleagues, so they themselves can start with more complex tasks early in their careers and move up the ladder more quickly, he stressed.
We could be seeing digital colleagues in the next year and a half
Globally, 82% of executives are thinking about introducing digital agents to their company in this timeframe. So they are asking, for example, how many agents can do different aspects of a person's job in their organisation? How many people are needed to manage a given number of AI agents? This is where a task-specific balance needs to be struck. For a well-defined process, one person can supervise up to 5 to 10 agents. The agent autonomously carries out the task, even makes decisions, but the process is ultimately controlled and evaluated by a human. In a strategic project, however, a much higher proportion of human involvement is required. The right proportion will be the key to success," said the Microsoft business unit director, analysing the possibilities.
The organisational chart will be replaced by the work chart, which means that work organisation will be built around project tasks rather than job titles and positions. Employees will be able to work in multiple roles, supported by AI colleagues. Competency-based organisation will come to the fore, who brings what skills to the task, building results-based, flexible teams by combining human and AI resources. AI means an extended knowledge base, where people can use their skills in different contexts. This leads not to chaos but to dynamic operations with good planning and organisation. The most important "metric" for leading companies is the degree of impact achieved, according to Rita Kamasinszki.
This also has a significant impact on recruitment, as AI affinity and openness to learning becomes an issue almost everywhere, as new entrants need to be able to adapt to these trends and changing work patterns.
In addition to the global conclusions, the business unit director also discussed the Hungarian situation. They see that there is openness to these technological solutions in Hungary, but in general Hungarian companies are still at the beginning of their journey. They acknowledge the importance of AI, there are pilot projects, but according to a BBJ study published in the spring, the majority of Hungarians have heard of AI, but its practical application is not yet common. 79% of respondents rated their own knowledge and practice as only 3.9 on a scale of one to ten. 27% of white-collar workers have only tried MI. However, the younger generation is very enthusiastic about AI - an advantage for companies when recruiting young people if they use such tools. 46% of Generation Z use AI tools on a daily basis, so if they enter the workforce, it will catalyse adoption.
Job seekers at risk
The expert says it is vital that companies invest in education on this topic, because they can only successfully use these technologies if they focus on employee training and change management. He reiterated the importance of vigilance and critical thinking in relation to the risks of using AI. The firm's recently published Cyber Signal Report shows that digital abuse/fraud in the job advertising market is on the rise. Fraudsters are creating fake job ads on various platforms, even conducting job interviews using AI in order to steal the sensitive data of candidates. Therefore, HR and job portals need to focus on security, even through multi-factor authentication. Employees need to be constantly trained on increased awareness and caution, as it is often the human being who is the weak link in these scams.
Photo by Rita Kamasinszki