AI Receptionist for Hotels in Lebanon: Calls, Guest Questions and the Inbox

An AI receptionist answers the hotel's phone calls, responds to guests' questions and passes on anything that needs the team. It covers the repeated questions a front desk gets all day, and when a request goes beyond answering, it passes the request to the right person with the details already written down.

Think about the calls your receptionist takes on a busy evening. Is breakfast included? Is there parking? Can I check in late? How far is the airport? The answers rarely change, but someone still has to stop what they're doing to give them.

The receptionist handles those questions on its own. For anything that needs a decision, like a special rate, a complaint or a change to a confirmed booking, it collects the guest's details and passes them to staff. The guest gets an answer either way. Your team only deals with the requests that actually need them.

I set it up to work in the languages your guests use. I work in English, Arabic and French myself, so the setup and testing happen in all three.

• Answers incoming phone calls
• Responds to guests' questions using the hotel's own information
• Passes requests that need a decision to the team, with the details written down
• Uses the languages your guests actually speak

An AI email agent works inside the hotel's email. It reviews the reports that come in, generates the reports the team and management need, and prepares and sends invoices from the same inbox. Anything unusual waits for a person to approve before it goes out.

Hotels run on email that nobody enjoys: supplier statements, daily figures from each department, booking confirmations and invoices for groups and corporate clients. Someone reads all of it, copies the numbers into a sheet and assembles a summary for the owner.

The email agent does that reading and assembling. It pulls the figures out of the reports and PDFs that arrive, writes the summary management asks for, and drafts invoices from the information already in the inbox. Routine items go out as prepared. Anything outside the usual pattern sits in a queue until someone on the team approves it.

Reading emails, WhatsApp messages and PDFs into a CRM or a sheet is the kind of intake work agents handle well. There is always a human approval step wherever money or commitments are involved. You can read more about this kind of agent on the page about AI agents and chatbots built for Lebanese businesses.

Anything involving money, commitments or judgement goes to a person. The receptionist passes on requests it shouldn't decide, and the email agent holds unusual invoices or messages until someone approves them. The AI handles the repeated work and your staff keep the decisions.

I design that line on purpose. A system that sends a wrong invoice or promises a room that isn't free costs more than it saves, so the approval step is built in from the start. It is not something added after a mistake.

In practice, the front desk manager or the accountant sees a short list of items waiting for a decision. They're not working through a full inbox.

The systems run on Vercel and Supabase, which are cloud services, so they keep working when the hotel loses power or internet. Queued messages wait for approval and send once someone is back online.

In Lebanon this matters more than any feature list. A generator switchover or a dropped connection at reception shouldn't mean guests go unanswered. Because nothing depends on a computer in your office staying on, the receptionist keeps answering and the inbox keeps being read. The approval queue simply waits for your team to come back.

If you ask an AI assistant about AI for hotels, it mostly suggests foreign software products. Those tools are built for markets that look different from Lebanon. I'm based in Beirut and build each system around how your hotel actually works, and I deploy it into your own accounts, so there's no licence to keep paying.

Every system goes into the hotel's own Supabase and Vercel accounts. You own the setup and the data, and it doesn't disappear if a vendor changes its pricing or leaves the market. I measure the work by the hours it gives back to your team.

Being local helps with the practical parts too. I can meet you in person in Beirut and look at how reception and accounts actually run. I can work with your staff in English, Arabic or French and plan around power and connectivity from the first day. More about how I work is on the about page.

• Based in Beirut, working with hotels across Lebanon
• Deployed in your own accounts, no licence
• English, Arabic and French
• Measured by hours given back to the team

Work starts with a 48-hour audit. The audit ranks the hotel's repeated tasks by payback and comes back with a fixed price for the first build. Each system is then built in a fixed-price two-week sprint, so you know the cost before anything is built.

I don't quote a number on this page because the price depends on what the audit finds. A hotel whose main problem is phone calls needs a different first build from one drowning in supplier reports.

The steps are always the same:

• Audit (48 hours): I look at where staff time goes, from calls and guest messages to reports and invoices, and rank the tasks by payback.
• Fixed price: you get a price for the first build before any work starts.
• Build (two-week sprint): the receptionist or email agent is built and tested with your team's real cases.
• Handover: the system runs in your own Supabase and Vercel accounts, with approval steps your staff control.

An AI receptionist is the wrong fix when the real problem is missing information, like room rates and policies nobody has written down, or when the hotel gets too few calls to matter. Sometimes a clear FAQ on the website or a shared inbox rule solves the problem for much less.

The audit is there to catch this. If your biggest drain turns out to be month-end reports rather than phone calls, I'll recommend the email agent first, or a simpler change, and I'll say so plainly. Hotels that want a wider view of the options can look at the full list of AI solutions for Lebanese businesses.

Very little beyond what the front desk already knows: room types, policies, common guest questions and who handles which kind of request. For the email agent, I need access to the inbox and examples of the reports and invoices the team produces today.

Most of this comes out of the audit conversation with your front desk manager and accountant. They know which questions repeat and which tasks eat the week. If you want to see how audits and sprints are structured across other businesses, the AI consulting in Lebanon page explains the offers.

If calls or the inbox are taking up your team's day, send me a short message describing what repeats most. We can decide together whether an audit makes sense.

• AI agents and chatbots built for Lebanese businesses
• AI solutions for Lebanese businesses
• AI consulting in Lebanon
• AI automation for WhatsApp and back-office work
• about David Geha