Table of Contents
Toggle
AI burnout is real, it is not a personal failing, and the fix is pace rather than more effort. We felt it ourselves this summer, and we work in AI every day. What we learned coming out of it is that the industry has moved to a stage where one small, well-chosen build beats another month of trying to keep up with everything.
What AI burnout looked like for us
Peter hit the wall first. By the time summer arrived he was so burnt out on AI that he could not do anything useful on his computer. His words on the episode were “completely useless,” and he meant it. This from someone who spends most of his working life talking about, testing, and building with these tools.
Kendel felt it coming into summer too. Skills had just launched in Claude. Peter was excited, finding use cases everywhere, and Kendel did not have the bandwidth to wrap her head around how to use them, let alone start playing. So she stopped. She spent time with her kids, read a stack of books she had never made time for, and used a different part of her brain for two months.
We are saying this because we think a lot of people in tourism are feeling it and keeping quiet. If you run a business, a destination, or a department, you already had a full job. Then AI showed up as one more thing to learn on top of it. The tools change every few months, but the conversation never stops. Nearly every conference has a session on it. It is tiring, and the tiredness is not a sign you are doing it wrong.
Two reasons AI fatigue hits tourism harder
Two things make this industry a hard place to keep up with AI.
The first is the calendar. Summer is when much of the tourism industry is busy hosting. It is a hard season to learn anything new, and it happens to be the season when the rest of the world publishes its biggest AI updates. Many operators come out of September behind on something they had no time to look at in July, and the feeling of being behind is its own kind of exhaustion.
The second is team size. A DMO with four staff or a tour company with two owners rarely has anybody whose job is “figure out AI.” It falls on the person who is also doing the marketing, the funder reporting, and the payroll. That person often does not have a spare afternoon. They have five minutes between other things, and most AI advice assumes they have a week.
There is research that matches what we saw in ourselves. In March 2026, Harvard Business Review published a study by researchers from Boston Consulting Group and UC Riverside on what they called “AI brain fry”: mental fatigue from using or supervising AI tools past a person’s cognitive capacity. People described mental fog, slower decisions, and more mistakes. Fatigue was lower when AI was used for routine work and when managers set realistic expectations rather than pushing for more output. That is a useful frame for a small tourism team. The goal of adopting AI was never to add a second job. If it feels like one, something about the pace is wrong.
What changed over the summer
The part that surprised us when we came back was that the pace eased a little. The models keep getting smarter, but the pressure to react to every release dropped off. The AI labs are talking openly about slowing some development for safety reasons, and estimates for how quickly organizations would adopt these tools have been revised downward because it is happening more slowly than predicted. Kendel put it plainly on the episode: for the two of us, that slowdown meant we finally had time to catch up on what had launched. For everyone listening, it means there is room to dig in without feeling behind.
The tools also got more useful without getting harder. The change of the last few months has been in what Peter calls the harness: skills, plugins, connectors, memory, and the ability for an AI tool to work through a multi-step task from start to finish rather than answering one question at a time. Claude Cowork and Claude Code are the examples we use most at Tourism AI Network. You can now hand a tool a repeatable process and have it run the process. That was not practical for a small team a year ago.
The requests coming into TAIN also changed. A year ago almost every conversation started with “I need to learn this.” This summer it turned into “can you build this for me?” One client did a training program with us, said it was great, and then said “I am running a business, I do not have time, can you just build me the things I need so I can run more efficiently?”
Training is still part of what we do, and there will always be a need for it. But that request tells us where the industry is heading. Some level of foundational understanding is there for many organizations. People have tried the tools and they know the work could be done better. What they want is someone to come in, help them rebuild the workflow, and hand it back working so they can learn on the fly.
Pacing, in practice
“Give yourself some grace” is a nice line, and Peter said it on the episode, but it needs a method behind it or it turns into permission to do nothing. Here is what pacing looks like for us this fall, and what we think a tourism team can borrow.
Five minutes and one question
When Kendel introduced a first-time user to Claude skills this summer, the first thing she had him do was ask Claude: look at the work I do, and can you suggest a skill that would be useful for me? That was the whole session. Five minutes, one prompt, and it opened a different way of thinking about the tool. If you are re-entering after a break or entering for the first time, start there. Ask the tool what it could do for the work you already do. Try one suggestion and stop for the day if you need to.
Pick one area and make it really good
This was Peter’s closing advice on the episode, and it runs against the instinct to try everything at once, which is exactly how you end up fried. Choose the one workflow that eats the most hours in your week. For a DMO, that is often a reporting cycle. For an operator it might be turning inquiries into quotes, or turning a season’s photos into a month of content. Fix that one thing before you look at the next.
Move from learning to doing
Peter’s line was “it’s less about learning now and it’s more about just doing.” A working automation in your own environment teaches you more about what AI can do for your organization than another webinar will. That is the whole premise of the Build Sprint: one team, one workflow, one working automation, built with you and left running after we leave. It exists because clients asked for it in almost those exact words.
Meet the tools where you are
Not everyone reading this is thinking about agents and orchestration. Plenty of organizations we work with are at the beginning: free accounts, no team plan, a few people experimenting. Nothing above should discourage you. Every TAIN engagement begins with an assessment of where your team is, what tools you already use, and what the workflows look like, so the recommendation fits your starting point rather than someone else’s. That is what consulting first means to us, and it is also what pacing means. You cannot pace a plan that was written for a different organization.
Let the machine do the routine work
The HBR research found fatigue was lowest when AI handled routine tasks, and that matches what we see. Giving AI the repeatable, boring, well-defined work, like the report that is due on the same day every month, the recap that follows every meeting, or the inquiry that gets the same four questions answered every time, leads to teams feeling energized by AI. Trying to use it for everything, including the things the team was already good at, is what leads to feeling drained.
What we are doing this fall
We are practicing what we just described.
Peter’s season is built around a short list: identify the most common tourism workflows we see across engagements, package them, and pressure test them with industry partners so they are easy to install and customize for a specific business. The beginning and the end of most of these workflows look the same across organizations, it is the middle that changes. Building the middle once, properly, is more useful to the industry than building it from scratch every time.
Kendel is starting several monthly embedded advisory engagements: working with teams on an ongoing basis, understanding the organization in depth, and building solutions alongside them rather than running a session and leaving. That is the slow, steady version of adoption, and it is the one that holds.
And we are both keeping the podcast going, because the conversation itself is part of pacing. Season 4 will bring more industry voices into the room, and every episode answers the same question: what did this person or organization build, and what did it take to get there. We think that is the question worth asking now, and it is a more useful question than “what launched this week.”
Where to start
If you came out of the summer feeling behind on AI, you are in good company, and the feeling is not a measure of how you are doing. Pick one workflow, ask one question, build one thing, then decide on the next.
Listen to the full conversation on The Tourism AI Podcast. Season 4, Episode 1 is available on Spotify, Apple Podcasts, and YouTube.
If you want to talk through where your team is, every engagement with Tourism AI Network starts with a conversation and a short pre-engagement survey. Email info@tourismainetwork.com to get started.