Show us the routine: Which AI job could a hotel switch on next Monday?

What goes in:

The reservations or sales inbox, plus live rates and availability from the PMS and the hotel's own meeting-room PDF and website.

What comes out:

For every group or meeting request that arrived after the front desk closed, a drafted quote sits in the reservations inbox. It covers rooms, rates and meeting space, and flags open questions such as dates, dietary needs or AV. The reservations or sales manager checks it, adjusts it and sends it.

When it runs:

Every time a request arrives, day and night, so the quotes are ready when the morning shift logs on. Planners often book the first hotel that replies, so a hotel that answers at 07:00 instead of 11:00 has a real head start.

An obvious one for me is checking reviews. Set a scheduled bot (GPT dot, or Grokbot, or Claude) to check the hotel's reviews every night, summarize them. And note the key good and bad points in a Google sheet.

What goes in: The previous shift's handover log and open guest requests, pulled from the systems the hotel already runs, with guest identifiers removed and an AI tool the property has already approved doing the reading. No new software, no new data, just the notes the team already writes.

What comes out: A "promises still to keep" brief on the duty manager's desk before every shift change: requests still open, commitments someone made to a guest, updates nobody recorded, each tied back to its source entry. The AI flags the gaps. The manager checks the facts, assigns an owner to every promise, and decides how each one gets answered.

When it runs: Before every handover, so no promise crosses a shift unowned. The outgoing team confirms what is truly closed; the incoming team hears what carries forward and why. Start with one shift and one department, review the misses together each week, and let the routine earn its place. The goal is fewer broken promises to guests, not fewer people keeping them.

Working on the assumption that the environment is secure......right?

What goes in: Consolidated Profile information for the coming days guests. This would include general profile content, booking information, historical behaviour data if available and any preference information, service requests/failures or notes about the customer or the stay/visit and, customer feedback.

What comes out: Asked to generate key insights and behavioural information about each guest. Suggests one thing that the operation could do to improve the specific guest experience or their revenue contribution.

When it runs: Daily in advance of arrival date.

Give it a whirl!

Improve guest communication 

What goes in: Ask the AI to read your entire website, your concierge mailbox (for tone of voice and context), and supplement it with additional information that only lives onsite (like your Wi-Fi password). Upload it to an AI messaging tool connected to WhatsApp, SMS and Booking.com chat. 

What comes out: The AI can answer almost any guest query pertaining to your hotel, reducing the workload for front desk teams and giving guest faster responses. Anything the AI is unsure of gets passed onto a human. 

When it runs: It runs instantly whenever anyone – guest or future guest – asks a question. And as a bonus, it can also answer in their native language. 

The panel agrees with a lot of research on AI and robotics, pointing out that repetitive and rule-based work should be automated first. However, in hospitality, it is often this repetitive work that makes up the operational work, and that is tricky to automate because it is physical (and therefore unable to be executed by software alone). It is also in this rule-based operational work that true hospitality experiences happen most often for the guests. Try to think back to your last memorable experience in a hotel. Chances are that this moment was made memorable by a waiter or a receptionist who paid a little extra attention to you, or said something that made you feel seen, through an interaction that is difficult to put into rules. While AI is attractive as a technology to execute repetitive, rule-based tasks, we should not forget that true hospitality often happens when these rules get interpreted flexibly by human staff.

What goes in: All guest feedback and guest-to-agent conversations from the last 24 hours, matched to the stay. In-stay, that means on-site feedback and guests’ chats with the AI agent, for example on WhatsApp. Post-stay, it means post-stay surveys and public reviews.

What comes out: A morning brief for the Front Office Manager. For guests still in house, it shows the tickets the agent opened overnight and flags any still unresolved before check-out. For guests who have left, it lists drafted replies to approve, edit or drop. The GM gets the top three recurring issues and any escalations to raise in the morning briefing.

When it runs: Daily at 7:00, before the morning briefing.

If you are a revenue manager, your morning routine is a good place to start. Every revenue manager has one that sets up the week ahead: On Monday you open the usual reports and check how the hotel is pacing, how pickup compares with last year, where the forecast sits against budget and how your rates look against the market. 

Immediately you can see the problem, and the opportunity. The reports are the same every week and the data sits in different systems, so a good part of the morning goes on collecting and analysing data before you get time to think about decision-making.

All of this can be delegated to AI. When you sit down, pace, pickup, forecast and market position are already pulled together with recommended actions.

It’s a sensible first job because the risk is low and your team knows these numbers well enough to check them against the sources to make sure it all adds up, building trust.

Automating guest conversations across channels

What goes in: Guest conversations across existing channels, including website chat, messaging, WhatsApp, email and even phone calls, supported by a centralized, AI-ready hotel knowledge base. Building and maintaining this structured knowledge used to be a major challenge, but today's AI solutions have largely overcome it, with minimal manual effort.

What comes out: AI automatically handles repetitive questions, providing immediate, consistent answers across channels. Complex requests or conversations requiring human judgment are escalated to hotel staff, freeing teams to focus on what really matters: selling and delivering exceptional guest experiences.

When it runs: 24/7 across digital channels, with phone calls initially covered outside regular service hours. Start with existing channels and expand progressively, without requiring a major technology project.