For boutique hotels, resorts, and independent hospitality groups.

Booking.com already takes 15–25% to find your guest.
Don’t let an AI travel agent take a second cut.

ChatGPT, Gemini, and Perplexity are already answering “where should I stay in Vienna” — the same question Google used to own. If your site isn’t built for those agents to read and recommend you directly, you don’t lose the booking. You lose it to whichever platform positions itself as the new middleman.

Most hospitality sites are built to look good in a carousel. I build sites that a machine can read, quote accurately, and route a booking through — while still being the best-looking option in the room for the humans who land on it.

Send a project brief

No RFP required. Tell me what you're running — PMS, channel manager, booking engine — and what isn't talking to what.

Why listen to me on this

I haven't built this exact stack for a hotel yet. Here's what I have built: a sync between an ERP and a storefront carrying several thousand SKUs — on a schedule, with a nightly reconciliation and no manual intervention. That's the identical architecture problem as keeping a booking engine, a channel manager, and a marketing site in agreement about rates and availability — just a different noun. Rooms instead of SKUs. Rate calendars instead of price lists.

If you want a hospitality logo on this page, I don't have one yet. If you want someone who's already solved the actual technical problem — real-time inventory sync, structured data an AI agent can parse, admin tools your staff can run without calling a developer — that's what's below.

The OTA problem you already know. The one you don't, yet.

You know the OTA math. Every booking that comes through Booking.com or Expedia hands over 15–25% before you've covered a single night's cost. You've probably already pushed direct bookings as hard as a static website and a “Book Direct and Save” banner can push them.

Here's the part that isn't on your radar yet: the next layer between you and your guest isn't a travel agent or an OTA. It's an AI system. Someone asks ChatGPT to plan a long weekend in Salzburg. Someone asks Gemini for “a quiet hotel near the Altstadt, not a chain.” Someone asks Perplexity to compare three properties on cancellation policy and breakfast.

Right now, most hospitality sites can’t answer any of those questions in a form the AI can use. The information exists — buried in a PDF policy page, a photo carousel with no alt text, a footer paragraph nobody wrote for a machine to parse. So the AI either skips you, or it answers from whatever OTA listing has the cleanest data. Either way, someone else gets credit for describing your property to the guest who was about to book it.

That's not a marketing problem. That's the same scoping gap heritage brands hit with their ERP: someone built you a beautiful site and never asked what happens when a system that isn't a browser needs to read it.

What it looks like when your site speaks the same language as your booking engine — and the AI agents reading it.

A guest asks an AI travel planner for a boutique property in your city with a specific amenity — quiet street, late checkout, a pet policy that actually says yes. Your site answers that question correctly, because the information was structured to be asked, not just written to be read.

Your rates and availability match everywhere, all the time — the channel manager, the booking engine, the website — because one system feeds all three, not because someone updated three places by hand before a rate change went live.

Your front desk stops fielding “your website said available and now you’re telling me it’s not.” The numbers agree because there’s one source of truth, not three spreadsheets pretending to be one.

You go into shoulder season knowing your seasonal rates, package pages, and availability calendar are correct — because your team updated them once, this morning, in an interface built for that job.

Direct bookings, built for how guests actually decide now.

  1. 01

    Real-time sync between your PMS and everything downstream

    Problem
    Your property management system knows your real rates and real availability. Your website usually finds out later — after a manual update, a plugin sync that runs on a schedule, or a spreadsheet someone forgot to touch this week.
    Cure
    I build a middleware layer — Cloudflare Workers — that reads from your PMS or channel manager's API continuously and writes to your website and its structured data in real time. No nightly batch job. No stale rate calendar.
    Proof
    This is the same pattern I run for a retail client's ERP-to-storefront sync — several thousand SKUs, carried across on a schedule with a nightly reconciliation, no manual step since the day it went live. Rooms and rates are a smaller, simpler version of the same problem.
    Offer
    If you're not sure whether your PMS or channel manager exposes an API, send a brief and I'll tell you within a week what's actually possible.
  2. 02

    A site built to be read by guests and by the AI agents advising them

    Problem
    Most hospitality sites are templated — same carousel, same generic “luxury experience” copy, same PDF buried three clicks deep for the policy an AI (or a guest) actually needs to answer a real question.
    Cure
    I build with the structured-data and entity-clarity approach I call AI-Ready — schema markup an AI can parse, policy and amenity information written in language a machine can quote accurately, and copy precise enough that when an AI describes your property, it describes it correctly. Underneath, a design that still has to win on sight, because humans book too.
    Proof
    asksj.com runs on exactly this approach — full entity graph, explicit machine-readable structure, no guesswork for anything reading the site that isn't a browser. It's the same method, aimed at a property instead of a studio.
    Offer
    Every project starts with an audit of what’s currently readable — by search, by AI, by a human on a phone — before I touch a design file.
  3. 03

    Back-office tools your staff will actually use

    Problem
    Rates, seasonal packages, and availability calendars usually live in a spreadsheet, a PMS interface nobody but the GM understands, or a request to “just ask the developer to change it.”
    Cure
    I build admin tooling on top of the same data layer powering the site and the booking flow — rate updates, package building, seasonal calendars — so your team changes it once and it's correct everywhere.
    Proof
    For a retail client, the back-office tools handle product curation, collection management, and content updates — run entirely by their team, no developer required for day-to-day changes. Same philosophy, different set of levers.
    Offer
    What gets built here depends on how your team actually operates day to day. The sync architecture is fixed; the tools on top of it aren't.

“Shouldn’t I hire someone who’s actually done a hotel before?”

Fair question. Here's the honest answer: I haven't shipped this exact stack for a hospitality client yet. I've shipped the exact architecture — real-time inventory sync, AI-legible structure, staff-run admin tools — for a client with a harder version of the same problem: thousands of SKUs instead of dozens of room types, a retail ERP built for the shops instead of a modern PMS API.

What you're actually hiring is the person who solves the sync-and-legibility problem, not a template with a “hospitality” label on it. I do the strategy, the build, and the maintenance myself — no account manager relaying your requests to a developer who's never spoken to you.

I've been independent since 2008. I incorporated in Vienna in 2018. Everything above is real work I can show you — it just isn't wearing a hotel's name yet.

What this costs

Projects in this category — booking-engine integration, site rebuild, back-office tooling — typically run between €40,000 and €150,000. Scope drives the number.

For a rough sense before a proper brief: a site rebuild without PMS integration starts around €10,000–25,000. Full-platform builds with real-time sync and custom admin tooling are €80,000 and up.

I don’t quote on a discovery call. I quote in writing, after I understand what you’re actually running.

Ready to talk about it?

Send a brief. Tell me your property, what you're running it on, and the specific thing that isn't talking to the rest. I'll read it and tell you within a week whether it's a fit and roughly what it would cost.

Send a project brief

I work with a small number of clients at a time. If the calendar is full, I'll say so.