Workers Spend 90 Minutes a Day With AI, and 2 in 5 Chats Aren't Human
BambooHR research shows employees now spend 90 minutes a day with AI and 2 in 5 conversations are with a machine — quietly eroding mentorship and workplace relationships.
By Nathan Brooks
4 min read

What's News
- Employees estimate they spend nearly 90 minutes per day interacting with AI, per forthcoming BambooHR research.
- About 2 in 5 of employees' conversational interactions are with AI rather than another person.
- A task that took an hour now takes 20 minutes with AI — a 67% time saving that usually fills with more tasks.
- Many employees say cross-generational mentorship is getting harder as younger staff rely more on AI.
Employees now spend nearly an hour and a half each day interacting with AI, and about 2 in 5 of their conversational exchanges happen with a machine rather than a colleague. That's the headline finding from forthcoming BambooHR research on how AI is reshaping the workday, and it lands at roughly nine full workweeks of AI interaction per employee per year.
The piece, published as an Entrepreneur contributor column built around the BambooHR data, starts from a simple scene. An employee has a question. A few years ago they'd message a colleague, walk to a desk or wait for the next team meeting. Today they ask AI and get an answer in seconds.
Why employees prefer the machine
The author — who reviewed the research — doesn't pretend the preference is irrational. AI is fast. It's always available. It doesn't look busy or make you feel you should already know the answer.
"You can ask the same question three different ways without testing anyone's patience."
Many employees in the research say they'd rather quietly ask AI than admit they need help. The column gets the appeal: asking a person carries social risk, and AI removes most of it. That makes people more independent, which has real value.
But the author keeps circling one doubt: "I wonder if we are also removing one of the simplest ways workplace relationships begin."
What did the interaction actually produce?
The core argument is that leaders have spent years stripping friction out of work, and AI is simply the latest and most powerful tool for that job. Poor systems, repetitive processes, pointless approvals and agenda-less meetings drain energy. Nobody should mourn those.
The problem is what sits inside the "inefficient" human interaction. The old desk-side conversation did more than deliver an answer. It showed a junior employee how an experienced person approached the problem. It surfaced context that was never written down. Sometimes, the author writes, "a small interaction becomes a relationship that matters much later."
The BambooHR findings back this up with a structural split:
- AI time concentrates on completing tasks, experimenting, refining prompts and correcting output.
- Coworker time includes execution, but far more of it goes to sharing knowledge, brainstorming, mentoring and building relationships.
Those relational outputs are hard to measure because they rarely produce an immediate deliverable. Their value shows up later — when a junior employee knows whom to call in a difficult moment, or has watched a senior colleague make a judgment call where the answer wasn't obvious.
The mentorship gap
This hits people early in their careers hardest. Most employees in the research say they've learned something valuable from a colleague at a different career stage. Many also believe cross-generational mentorship is getting harder as younger employees lean more heavily on AI.
The tension the author identifies: "We value what we learn from one another while adopting tools that let us skip some of the awkwardness, delay and vulnerability that often start the learning process."
AI can give a good answer — maybe a better explanation than a rushed colleague. But professional judgment, the author argues, is rarely just the answer. It's knowing when the technically correct answer won't work in this company, which trade-off matters most, or recognizing a pattern because you've seen a version of the problem before. That's learned by watching other people think.
The efficiency dividend nobody is spending
What does this change for managers? The column's sharpest practical point concerns reclaimed time.
"We have spent a great deal of time deciding what AI should do. We have spent less time deciding what people should be freer to do because of it."
The default answer in most organizations is more work. A task that took an hour now takes 20 minutes, and the surplus fills with another task, another message, another expectation. The efficiency vanishes into a fuller calendar — a 67% time saving per task that no one ever banks.
The author's alternative: if AI hands a manager back an hour, some of it should go to coaching. If an experienced employee gains capacity, they should spend part of it explaining how they reached a decision, not just sending the answer. That's work too, and it likely becomes more valuable as transactional work gets cheaper.
The column closes without nostalgia. Some friction should disappear, and the employee who asks AI first may still be right. But AI will let us need each other less for everyday work — and the practical question for every leadership team is what fills the space that creates.
Original: eu.entrepreneur.com
More from Nathan Brooks
Show full bio
News editor covering regulation, policy and their impact on entrepreneurship, startups and venture funding at Business Bearings.
2,349 articles