AI Is Changing Work. What Are We Making Room For?

Setting up the phone line for my new consulting business sounded simple: pick a number, record a greeting, carry on. Then came the question of what callers would actually hear. Was the automated introduction warm enough? Did it sound professional? Did it sound like a real person might answer the phone eventually?

For a brief moment, I had become the executive producer of a fifteen second telephone greeting. The awards ceremony, I assume, is forthcoming.

Technology made the mechanics easier. It could not decide what kind of welcome I wanted to give. The stakes in a phone greeting are low; the stakes in an employment decision are not. Yet the underlying distinction matters in both places: a tool can accelerate the work, while people remain responsible for what the work is meant to accomplish.

The productivity question has a second half

Deloitte’s 2026 Global Human Capital Trends research found that 85% of leaders said it was critical to build the capacity to adapt at the speed required today, while only 7% said they were leading in helping their workforce continuously grow and adapt. It also found that only 6% of leaders reported progress designing interactions between people and AI. Source: Deloitte, March 2026.

That gap deserves attention. After years working with leaders on the people side of change, I have learned that a faster process is rarely the whole answer. A tool may take a four hour task down to twenty minutes. Then comes the leadership question: What happens to the other three hours and forty minutes?

Perhaps people get to spend more time with customers, coach their teams or solve problems that have sat untouched for years. Perhaps the organization simply adds another four hours of tasks and congratulates itself on reducing workload. Both outcomes are possible. Neither is decided by the software.

Let the tool assist. Keep a person accountable.

Microsoft describes emerging organizations as “AI-operated but human-led.” It is a useful phrase if we give the second half real meaning. Source: Microsoft, 2025 Work Trend Index.

In HR, AI can help organize information, summarize themes, create a first draft or identify questions worth investigating. A person must still check the facts, understand context, protect privacy, consider fairness and own the decision.

A polished draft of a performance letter can be produced quickly. That does not tell us whether the employee received timely feedback, whether the expectations were reasonable or whether the letter is accurate. I have seen enough HR processes to know that a beautifully formatted document is not evidence of a fair process. Efficiency cannot repair one.

This matters far beyond HR. If an automated recommendation affects someone’s work, opportunity or livelihood, a leader should be able to explain what information was used, how it was checked and who is responsible for the result. “The system suggested it” will be a very unsatisfying answer to the person affected.

Four questions worth asking before rollout

What problem are we solving? “We need to use AI” is an intention, not a problem statement. Name the work that is too slow, repetitive or inconsistent.

Which moments should remain human? Identify where a person needs to listen, weigh context, explain a choice or be visibly accountable.

How will people learn the new work? Teach employees to check outputs, question recommendations, protect information and recognize when a tool is out of its depth.

How will we know the work got better? Measure quality, experience and trust alongside time saved. An impressive speed gain that leaves employees anxious and customers confused is an incomplete result.

There is reason to be hopeful. Workplaces contain a remarkable amount of drudgery that good tools could remove. I would welcome more room for listening, thinking, teaching and solving difficult problems with other people.

The next chapter of work will be shaped by what leaders do with that new room. We can fill it with more noise. Or we can give people more opportunity to do the work that calls for care, judgment and imagination. I know which one I would rather help build.

When a tool gives time back, decide deliberately what deserves that time.

Original writing and illustrations for Work, Considered., the Work, Advised. blog. Research sources are linked at the relevant claims. Personal openings draw on Daniel's own background and experiences; no confidential workplace examples are used.