Learning Center AI
Leadership & Operations Oct 28, 2026 · 4 min By

Your AI Time Savings Are Already Spoken For

AI will save your center time, but it won't give you space unless someone decides what the recovered hours are for. Lessons from Daniel Sih and Paul Matthews' Wise AI, plus three questions for your next staff meeting.

Ask around your center and you will probably hear some version of the same thing. AI has made the email faster. The report drafts come together quicker. The tutor training materials that used to eat a Friday afternoon now take an hour.

So where did the hour go?

Most of us cannot say. It got absorbed into something. A meeting that expanded. Another report someone asked for. Two more initiatives that seemed manageable because, well, we have the capacity now.

This is the problem Daniel Sih and Paul Matthews take on in their book Wise AI. Their argument is simple and uncomfortable. AI will save you time. It will not give you space. Those are different things, and the second one does not happen on its own.

Why the hours disappear

The pattern is older than AI. Sih and Matthews point back to Thorstein Veblen's work on the leisure class to make the point. When a culture treats leisure as a mark of status, efficiency creates free time. When a culture treats busyness as a mark of status, efficiency creates more work.

Look at how we talk about our calendars in this field. An empty afternoon does not read as well-managed. It reads as underutilized. Nobody says so out loud, but the person with back-to-back weeks is the one we describe as committed.

Under those conditions, every minute AI gives back is an invitation to take on something else. Parkinson's Law does the rest.

The workplace evidence backs this up. Sih and Matthews cite research showing that employees using AI worked faster, took on a wider range of tasks, and stretched work into more hours of the day, often without anyone asking them to. Microsoft's own workplace research found that a majority of workers struggle to find time for their actual job because so much of the day goes to talking about work.

None of that is a failure of the technology. It is what happens when nobody decides in advance what the recovered time is for.

Where this hits a learning center

The fastest AI wins in our work are in the administrative layer. Scheduling communications. Session note templates. Tutor training materials. End-of-term reports. Survey summaries. The work that fragments attention and fills a day without ever being the reason you took this job.

AI genuinely compresses that work. I have seen it in my own week.

The question is what happens next, and it is a supervisory question rather than a technical one. Those recovered hours can fund more programming, which is the default because programming is visible and countable. Or they can go back into the thing that administrative load has been eating for years: time in the center, time with tutors, time with students.

Both are defensible. What is not defensible is letting the decision make itself, because when it makes itself, the answer is always more work.

The part that should worry us more

Sih and Matthews list ten skills for working with AI well. Nine are about using it better. The tenth is about not using it.

Keep doing hard things without AI, they argue, or the capacities you need to use it well will erode.

That one belongs in tutor training, and it belongs in our conversations with students. The effort of retrieving and practicing material is what produces the learning. A student who hands that effort to a chatbot has removed the exact thing that makes studying work. They will feel more prepared and be less prepared, which is the worst possible combination going into an exam.

The same risk applies to our staff. A tutor who outsources session prep stops building the content fluency that made them good in the room.

Three questions worth putting on an agenda

If you want something concrete out of this, try these at your next staff meeting.

What did AI actually save us this month, and where did it go? Not a guess. Pick one process you changed and trace the hours.

What are we protecting? Name the thing the recovered time is supposed to fund, before it gets spent. Write it down somewhere other people can see.

What should we still be doing by hand? Both for us and for the students we work with. The answer is not nothing.

One caution while you are at it

Sih and Matthews have a phrase I keep coming back to: data incontinence. It means feeding sensitive information into AI tools without thinking about where it goes.

Session notes. Student names. Advising details. The convenience of pasting a note into a chatbot and asking for a summary collides directly with FERPA, and it collides quietly, because nothing visibly goes wrong.

If your center is using AI for administrative work, and most are, this needs to be a stated expectation rather than an assumption.

Sources

AI use statement

Sources come from my own reading, plus articles flagged by Elvis, an AI agent I built in Muse (Meta, 2026). I save the ones worth keeping with a note on why they matter for learning centers. Claude (Anthropic, 2026) helps me review those notes, choose topics, and outline and draft the article and its prompts. I revise every draft, check claims against the sources, and decide what gets published.

I tested the prompts in ChatGPT (OpenAI, 2026), Claude (Anthropic, 2026), Copilot (Microsoft, 2026), Gemini (Google, 2026) and Grok (xAI, 2026).

Sources come from my own reading, plus articles flagged by Elvis, an AI agent I built in Muse (Meta, 2026). I save the ones worth keeping with a note on why they matter for learning centers. Claude (Anthropic, 2026) helps me review those notes, choose topics, and outline and draft the article and its prompts. I revise every draft, check claims against the sources, and decide what gets published.

I tested the prompts in ChatGPT (OpenAI, 2026), Claude (Anthropic, 2026), Copilot (Microsoft, 2026), Gemini (Google, 2026) and Grok (xAI, 2026).

References

About the author

[[shortcode.authors.mark-woolwine]]

← Back to The Feed