How Is AI Changing the Way Teams Work Together?
Last updated: 17 October 2026 · By Madi Hanc, PhD, ThinkWorkPlay
Direct answer: Recent research suggests AI may be changing how teams work together in three ways. Frequent AI users report collaborating with colleagues more often, meetings appear to be shifting towards joint problem-solving and decision-making, and many workers say their overall volume of work is growing. There are early signs of a risk alongside the benefit: when people use AI in place of conversations with colleagues, trust and coordination may weaken. The evidence is still emerging and mostly self-reported, drawn from this week's Jabra WorkLife 2026 survey, new research in Harvard Business Review and PwC's latest workforce survey.
The week's throughline: new patterns of work are forming around AI
Working alongside AI is commonly assumed to save time, to leave people more isolated from their colleagues and to make meetings less necessary. This week's research looked at what happens to the shape of the working day, and to the relationships inside teams, once AI becomes routine.
In a plot twist, the findings challenge each of these assumptions: workloads appear to be getting heavier, frequent AI users report collaborating more, and meetings seem to be taking on new purposes. Together these point to new patterns of work, and there is little sign that most organisations are designing them deliberately.
The study: Jabra WorkLife 2026
WorkLife 2026: The New Shape of Work, published on 13 October by the Jabra Research Institute, surveyed more than 6,000 knowledge workers in France, Germany, India, Poland, the UK and the US across almost 200 variables; according to the press release, Toluna carried out the fieldwork in May 2026 (Jabra Research Institute, 2026; Jabra press release, 13 Oct 2026). The findings deserve some care: the study is vendor-commissioned (Jabra sells audio technology), self-reported and cross-sectional, so it can show associations but cannot prove cause and effect; it covers knowledge workers only; and even the full report gives no per-market sample sizes, weighting or numeric threshold for its "AI Superusers", the roughly 15% of workers who combine three signals Jabra calls Integration, Organisational Enablement and Transformation.
Pattern 1: workloads appear to be growing alongside AI
Jabra calls this a productivity paradox. AI may be giving workers time back, yet 85% say they do more of at least one kind of work than a year ago. The biggest increases are in strategic planning and analysis (51%), creative work (50%) and, notably, administrative tasks (49%). Only 35% say any part of their work shrank at all (Jabra Research Institute, 2026).
PwC's Global Workforce Hopes and Fears Survey 2026 of 49,364 workers adds context: 27% cite fatigue and burnout as a key limit to their productivity, and among workers affected by change, 34% agree they experienced more change in the last year than in previous years (PwC, 2026).
So what: saved time seems likely to be absorbed unless someone decides where it goes. Without deliberate choices, AI could quietly raise the ceiling on what a day is expected to hold.
Pattern 2: heavy AI users report collaborating more
Overall, 42% of workers say they collaborate with colleagues more than a year ago, against 13% who collaborate less (Jabra Research Institute, 2026). Some 64% of frequent AI users say they work more with colleagues than a year ago, compared with 25% of those who use AI less frequently (Jabra press release, 13 Oct 2026).
Connection follows the same pattern. Across the whole sample, 39% say AI has strengthened their sense of connection with colleagues and 14% say it has weakened it; loneliness and disconnection, both at around 6%, are among the least reported emotions in the study (Jabra Research Institute, 2026).
So what: in this sample, the fear that AI isolates people is not borne out for the heaviest users. Given the study design, though, we cannot say whether AI drives the collaboration or collaborative people drive AI use.
Pattern 3: meetings may be changing purpose
As AI takes on more drafting, summarising and status updates, meetings appear to be shifting towards work that still needs people in the room. In Jabra's data, 40% say most of their meeting time goes on problem-solving and 39% on decision-making (Jabra Research Institute, 2026).
Jabra also reports that people who regularly speak to AI, rather than mainly typing, are 24% less likely to link work with stress and 23% less likely to report exhaustion (Jabra Research Institute, 2026). Given Jabra's commercial interest in voice, this is the finding I would most like to see independently replicated.
So what: if meetings are becoming the place for judgement and decisions, they deserve better design: fewer attendees, clearer decision rights and preparation done in advance.
The counterpoint: "relationslipping"
A Harvard Business Review article published on 15 October by researchers from BetterUp and Stanford University's Jeffrey Hancock adds an important caution (HBR, 15 Oct 2026). They report that treating AI as a teammate reduced trust among team members and weakened their ability to coordinate, a pattern they call "relationslipping".
It happens in two ways. When people skip a conversation for a quick AI answer, they stop updating their sense of who knows what. When colleagues offload work to AI, their thinking becomes less visible to the team. The authors stress that AI is not inherently harmful to teams, and recommend using AI to prepare for conversations rather than replace them. (The public article does not describe the study's sample or design in detail.)
The evidence on human-AI teams: learning may be the missing ingredient
A new paper in PNAS by Julian Berger, Stefan Herzog and colleagues at the Max Planck Institute for Human Development and partner institutions adds peer-reviewed evidence on how people and AI work together (Berger et al., PNAS, 2026).
The authors reanalysed 74 studies from a 2024 review which had found that human-AI pairs, on average, performed worse than the better partner working alone. Only 10 of those studies gave participants any feedback on their decisions. Pairs given both feedback and AI explanations tended to outperform the better partner alone, while explanations without feedback were linked to clearly worse results. The analysis is correlational, and the authors note it is still unclear whether the findings extend to generative AI.
So what: read together, this week's research suggests AI may strengthen collaboration when it feeds conversations between people and gives them a chance to learn, and erode it when it substitutes for them. These ideas still need testing with stronger, independent data.
Who is missing out on the new patterns
The new patterns do not appear to be reaching everyone. In Jabra's data, 32% say their work has changed little or not at all, and this group reports the lowest satisfaction, preparedness and career optimism. Employees with an AI-capable manager are 2.4 times more likely to use AI daily (65% versus 27%) (Jabra Research Institute, 2026).
What leaders should do next
Decide where saved time goes. Name it explicitly: deeper work, development, recovery or flexibility. Otherwise it is likely to fill with more tasks.
Redesign meetings for judgement. Use AI for pre-reads and summaries, and keep meeting time for problem-solving and decisions.
Protect the "who knows what" map. Encourage people to share how they used AI and to check with colleagues before relying on an AI answer.
Watch the workload signals. Track burnout and after-hours work alongside adoption metrics.
Build in feedback loops. Give people ways to check AI outputs against outcomes, so trust in AI is earned.
FAQ
Does AI make employees collaborate less?
Early evidence suggests not, at least for frequent users. In the Jabra WorkLife 2026 survey, 64% of frequent AI users say they collaborate more with colleagues than a year ago, compared with 25% of those who use AI less frequently. HBR research warns, however, that using AI in place of colleagues may reduce trust and coordination.
Why isn't AI reducing workload?
Survey evidence suggests time saved by AI is being absorbed by new and expanded tasks. Jabra found 85% of knowledge workers do more of at least one type of work than a year ago, including more administrative work (49%).
What is "relationslipping"?
A term coined by BetterUp and Stanford researchers in Harvard Business Review (October 2026) for the weakening of working relationships when people turn to AI instead of their colleagues.
Comments