Insights
How do you make time for AI when you're already burned out?
You don't find it. That's the honest answer, and it's why the question feels unanswerable — nobody finds a free week inside a schedule that was overbooked back in March. What you can do is make one deliberate opening on purpose. One day. Closed schedule, no demos, nothing to learn, spent looking honestly at where your practice's hours actually go. That pause isn't the preparation for the work. It is the work. Almost everything artificial intelligence can honestly do for a healthcare practice depends on it, and almost nobody takes it, because taking it feels like the most expensive thing on the list.
The tiredness is real, and it is not a character flaw
Let's start where you actually are, because skipping this part is how most technology conversations lose practitioners in the first thirty seconds. The AMA's Organizational Biopsy — nearly 19,000 physician responses from 106 health systems and organizations across 38 states — found that 41.9% of physicians reported at least one symptom of burnout in 2025. That number is genuinely improving; it was 48.2% in 2023, and the people working on this deserve credit for the difference. It is also still two out of five human beings. And when the AMA asked what was driving it, the answers weren't hard cases or long hours in the treatment room. They were ineffective EHR systems, inadequate staffing, and excessive administrative tasks. Nothing on that list is a clinical skill. Nothing on that list gets better by caring more, and most of us have already tried caring more. The work wearing people down is, to a startling degree, not the work anyone trained for — and that holds for the hygienist charting through lunch and the coordinator on hold with an insurer just as much as for the practitioner who gets home, eats, and opens the laptop again at nine.
Why the pause feels more expensive than the exhaustion
Here is the arithmetic that runs quietly in the back of your head when someone suggests you set a day aside. Closing the schedule has a number. You know that number to the dollar — a day of production, a waiting list to re-contact, a team you're paying either way, and the particular dread of rebuilding a calendar that took six weeks to fill. Burnout, by contrast, has no line item. It never shows up in a monthly report. One side of the ledger is precise and the other is invisible, and precise always wins.
Except the invisible side has been measured. Shasha Han and colleagues, writing in the Annals of Internal Medicine in 2019, estimated that physician turnover and reduced clinical hours attributable to burnout cost the U.S. health system roughly $4.6 billion a year — and about $7,600 per employed physician, per year, at the level of a single organization. That figure doesn't include the hygienist who gave notice in March, or the eleven weeks the front desk sat half-staffed, or the quiet cost of a practitioner who went down to four days because five had become unsurvivable. You have already paid for the day. You paid for it in installments, without a receipt.
Then there's the second sentence we hear, and we hear it almost word for word: we're not business people, so we can't tell whether any of this is worth it. That's fair, and it's also the part we'd most like to put down. You don't need to become a business person to do this well. What the room needs is someone who can describe one ordinary Tuesday in granular, unflattering detail — what got touched twice, what got typed in by hand after it had already been written down somewhere else, what made someone sigh. That's clinical knowledge, not business knowledge, and you are the only person in the building who has it. Anyone who arrives for that day with a platform already chosen and no questions to ask is selling. The right kind of help spends the morning listening and doesn't open a laptop until after lunch.
From the field
A practice we work with agreed to this and then negotiated us down — we asked for a day, they offered an afternoon, and we took the afternoon, because an afternoon that happens beats a day that keeps getting moved. The first thing we did was not demo anything. We asked everyone in the room to write down every task from the previous workday that hadn't needed a human being. The list ran four pages. Then we picked one path — a new patient's first phone call through to a confirmed, verified appointment — and walked it out loud on a whiteboard, every handoff named. Two of the steps existed because of a software limitation that had been fixed two years earlier. Nobody had noticed, because nobody had ever been given a quiet afternoon in which to notice. The honest rest of it: they lost that afternoon's production and never made it back. The first tool they picked was the wrong one and they dropped it in week three. The time didn't really show up until around week six, and when it came it arrived as fifteen minutes here and twenty there, not as a free afternoon. They'd do it again. The relief, they told us, started on the day itself — before a single thing was automated, somebody had finally written the weight down.
10.8 minutes
Saved per workday by clinicians using an ambient AI scribe, measured 30 days in — a 2025 JAMA Network Open study across 263 clinicians at six health systems. Burnout in that group fell from 51.9% to 38.8%. The honest caveats: no control group, participants volunteered (so they were already friendly to the technology), and the surveys weren't anonymous. Eleven minutes a day is about an hour a week. That is a door opening, not a renovation — and an hour a week is still an hour you didn't have.
What one day actually looks like
When we run this day — and running it is most of what Vision Optimized™ actually does as an AI education and consulting company — it has four movements, and only one of them involves technology. The morning begins with naming the friction. Everyone, clinical and administrative, writes down every task from the previous workday that did not need a human. Not "what should we automate," which only invites people to guess at software they've never used. Just: what did you do yesterday that a competent system could have done, and what did you do twice? This sounds too simple to be the intervention. It isn't. In 2018, leaders at a health system in Hawaii published a short piece in the New England Journal of Medicine about a program they called Getting Rid of Stupid Stuff — they asked every employee to flag documentation in the EHR that was poorly designed, unnecessary, or just plain dumb. They got about 200 suggestions and went on to eliminate dozens of documentation processes outright. No artificial intelligence was involved in any of it. The asking was the active ingredient.
The second movement is following one path all the way through. Pick the single heaviest journey in your practice — a first phone call to a confirmed, insurance-verified appointment, or the end of a visit to a closed, billed, filed note — and walk it out loud in front of everyone, naming every handoff and every person who touches it. Nearly every time, somebody says the sentence that pays for the day: wait, why do we still do that? Only then, in the afternoon, do you choose one thing. One workflow, one tool, one owner with an actual name, one number you'll look at in thirty days. Not a platform, not a suite, not a transformation. One. Practices that try to adopt five things adopt zero, and the honest reason is that they ran out of attention, not money. The AMA's 2026 survey on augmented intelligence, of nearly 1,700 physicians, found 81% now using AI professionally — more than double 2023 — with 70% seeing it as a way to automate precisely the tasks that feed burnout. Appetite is no longer the bottleneck. Choosing narrowly is. And the fourth movement, the one almost everybody skips, comes at the end of the day.
Decide what the hour is for before you earn it
If you free up five hours a week without deciding in advance what those hours are for, you already know what happens to them. They fill. Two more patients, a meeting that didn't exist before, the same volume of messages spread comfortably into the gap. Recovered time does not stay recovered by default — it gets quietly reabsorbed, and then everyone concludes the technology didn't work, when in fact it worked fine and the practice spent the proceeds without noticing. So you decide on the day, in writing, while the time is still hypothetical and nobody is fighting over it yet.
Some of what practices choose is straightforwardly financial, and should be: work the denials properly instead of writing them off, open the morning block you've been too thin to staff, stop paying overtime to finish charts that should have closed at five. Profitability is a legitimate thing to want — a practice with margin is a practice that can keep its people and say no to the contracts that are grinding it down. But some of what the time is for isn't financial at all, and we want to say plainly that this is allowed. A Friday that genuinely ends at four. Being at the game. The teaching you keep deferring. The clinical interest you shelved years ago and still think about. The book, the garden, the second practice, your kid's last season before they leave. When we ask practitioners what they would actually do with a recovered afternoon, almost no one's first answer is "see more patients" — and almost everyone's written plan, before we ask, is to see more patients. Say the true answer out loud and put it on the page. It's the one part of this that nothing else can hand back to you.
There's research behind why a day shaped like this outperforms the usual remedies. In 2017, Panagioti and colleagues published a meta-analysis in JAMA Internal Medicine pooling 20 controlled interventions across roughly 1,550 physicians. The benefits overall were modest, but the strongest evidence belonged to organization-directed changes — changes to how the work itself flows — rather than to interventions aimed at individual resilience. They also found that the organization-directed kind were rare. A day where a practice redraws a workflow, names an owner and sets a date is an organization-directed intervention. A wellness webinar at seven in the evening, after the notes, is not. Only one of those two has the evidence on its side, and it's the one that costs you a day.
We're not going to tell you that one day fixes burnout. It doesn't. Burnout has causes that live well above your practice — reimbursement, staffing shortages, documentation requirements nobody in your building wrote. What one day does is smaller and more honest than that: it turns a vague, heavy dread about technology into a short list of specific, boring, fixable things, each with a name beside it and a date to check it. That alone changes how a team feels on the drive home, usually before anything has been automated at all. So if you take one thing from this, let it be that you are not behind. You're tired, which is a different problem, and it deserves a different first step than a software demo. Close the schedule once. Bring everyone. Write the weight down. Then pick one thing — and we'll take up how to choose that one thing, and how to know in thirty days whether it worked, in the next piece. If you'd rather not run the day alone, that part is the work we do.