AI training for employees: how hotels close the skills gap

Most hotels now hire graduates who used AI throughout their studies, yet few have a plan for teaching them how AI is used at work. That gap is where AI training for employees earns its place: not as a one-off workshop, but as part of how a team picks its tools, onboards new people and protects time for practice.

The lessons below come from a conversation about the Future Talent Pipeline report, a study of generative AI in talent management for hotel sales, marketing and revenue management. Two people who worked on it shared what the data and their client work show: Michael Goldrich, an AI advisor focused on hospitality, and Mike Gamble, a CEO in executive recruiting for travel, tourism, hospitality and events.

Confidence with AI is not the same as competence

Students arrive comfortable with AI. In Michael Goldrich's experience, most are self taught: universities do not show them how to use the tools, they simply use them. That makes them confident, but their competence may lag behind. Most of their use is writing help, such as drafting reports.

“However, when they get to the hotels, it's really all about numbers and analysis.”

Michael Goldrich, 4:14

The risk he describes is a false sense of authority. Someone who never had to question an AI answer at university may not question it at a hotel either. A second friction point is the tool: a graduate used to ChatGPT or Claude may join a company that only allows Copilot and feel constrained. Both are reasons to set expectations on day one: how this company uses AI, with which tool, and where human judgment checks the result.

Where the preparation gap is widest

The report sampled students in all four years. Seniors used AI about twice as much as juniors, and the gap between how confident students feel and how prepared they are was wider early in their studies.

Michael Goldrich links that to capstone projects, which require analysis and full project work instead of writing alone. Students who do that kind of realistic simulation feel more prepared, so his suggestion is to bring real work simulations in earlier. Hotels can support that through workforce readiness programs.

For a hotel, the same logic applies to the first weeks of a new role:

• Give new hires real analysis tasks with AI, not only writing tasks.
• Show how the company's own tool behaves compared with what they used before.
• Make checking and questioning the output an explicit part of the job.
• Pair them with a mentor who has been with the company for a couple of years.

That last point comes from Mike Gamble's client work. Some clients match graduates with younger mentors instead of senior ones: colleagues with a similar mindset who already understand how the company and brand get work done, on the human side and with AI.

Personal AI agents, shadow AI and the tool gap

The pace of change makes the tool gap bigger. Michael Goldrich points to the always-on personal agents that large tech companies have announced, which know a user's work and intent. A student who studies with such an agent and then joins a hotel that forbids it for security reasons will, in his view, either use shadow agents secretly on personal accounts or feel the company is not a good fit. His assessment is that hotels are not there yet.

The four T's: a framework for AI training for employees

Why does AI often stay in small pockets of a company? Michael Goldrich showed hotel management companies what ChatGPT could do soon after launch, and a year later nothing had changed. Lasting adoption starts with embracing change in the workplace. Where things did change, four conditions were in place. He calls them the four T's:

• Tone from the top. The CEO says this is important and everyone gets on board. Without that, people experiment in small areas and nothing is unified.
• Tools. Everyone uses the same tool, and onboarding and learning programs are built around it. He observes that many companies already have Copilot, yet hardly anyone really uses it.
• Training. Explain simply what the chosen tool can do. Mike Gamble values this for level setting.
• Time. The hardest one. People need to use the tool every day and need time set aside for it.

Time is hard because adoption has a curve. Michael Goldrich compares it to a new manager learning to delegate: at first it feels faster to do the work yourself. AI training for employees has to allow for that dip.

“If you spend the time and you start to learn it and you learn how to adopt everything, then that's when you get the benefit.”

Michael Goldrich, 19:16

Onboarding as a way to attract talent

A strong onboarding program does more than prepare one person. The view shared in the conversation is that the next generation is less likely to stay with one employer for decades, and that they talk to each other. A graduate who gets excellent onboarding and training on all the tools will tell friends and post about it, which attracts better talent. A company that says "here you go, good luck" gets talked about too.

Mike Gamble sees the same in recruiting. Culture is easy to research now, and onboarding, training and how a company uses AI are no longer a mystery to candidates. Winning the best talent is one step; in his view, the key is then to retain them. He looks back on the long training programs hotel companies once ran for graduates.

“Our companies took a lot of time to make sure we were coming out of college, and they still took a lot of time to make sure that we understood the culture and how things worked.”

Mike Gamble, 9:02

His conclusion is that the industry needs to get back to some of that. Modern AI training for employees can be the place where that investment returns.

Transparency, time and the human skills that still decide

Two more lessons apply to any team building AI skills. First, be open about monitoring. The report flagged transparency about how a company uses AI to monitor performance as important, and the conversation called secret monitoring a possible ticking time bomb. Michael Goldrich's advice is to tell new hires plainly that work devices are monitored.

Second, treat AI as a way to create thinking time. As people move up, they are paid to think but have less time for it. Michael Goldrich calls the technology a possible time machine, provided you invest the time first. The same goal drives artificial intelligence in hotels: freeing staff for guests.

None of this replaces people skills. From entry level roles to CEO searches, Mike Gamble says, EQ, soft skills and instinct still decide, especially in commercial roles. You can watch the full conversation or explore more episodes in the HSMAI Studio series.

Frequently asked questions

What should AI training for employees include in a hotel?

According to AI advisor Michael Goldrich, it should cover the one tool the company has chosen, explained simply, plus practice with real analysis tasks rather than writing alone. It also needs protected time to use the tool every day, because people get slower before they get faster.

Are hospitality graduates ready to use AI at work?

They are confident users, but confidence runs ahead of competence. Many students teach themselves and mostly use AI to write reports, while hotel roles in sales, marketing and revenue lean on numbers and analysis. The gap narrows in the final year, when project work becomes more realistic.

What is shadow AI in the workplace?

Shadow AI is when employees use their own AI tools or personal agents on private accounts because the employer blocks them or offers nothing comparable. Michael Goldrich expects it when graduates who studied with always-on personal agents join a hotel that forbids them for security reasons.

Should hotels tell new hires how AI monitors their work?

Yes. The conversation called transparency about AI monitoring essential and warned that secretly using AI to monitor performance could become a ticking time bomb. New hires should expect that anything done on a work device is monitored, and employers should say so openly.