Luxembourg, Luxembourg, Luxembourg
9 k abonnés
+ de 500 relations
9 k abonnés
+ de 500 relations
À propos
I help hotels remove operational overload using AI.
Most owners and operators I…
Most owners and operators I…
Activité
9 k abonnés
-
Piers Thackray a partagé ceciLast week, I refused to onboard a new hotel client. The owner wanted us to build all his AI workflows. I told him, "Try Claude or Claude Code for two weeks first. Then we'll talk." He looked confused. Here is why I said it. If your vendor builds your AI, and one day they walk away, what happens to the workflows your team has come to rely on? That fear is real, and it is reasonable. Most vendors will not say it out loud. Some build AI as a black box on purpose, so you stay tied to them forever. That is not the partnership I want. How I think an AI vendor should work with you: ▪ Your team understands what the AI does, not just sees the output ▪ Your hotel owns the accounts, the data, and the workflows ▪ If you ever want to switch providers, you can ▪ Training the team is part of the project, not an upsell If a vendor does not push for that from day one, red flag. How you start learning AI yourself, no technical background needed: 1. Open Claude (or any AI you trust) 2. Ask: "How do I build an AI agent that does X?" 3. Follow the steps it gives you 4. When you get stuck, take a screenshot, paste it, ask again 5. Keep going until it works That is it. No course. No bootcamp. No consultant. The hotels where the owner has tried this themselves are night and day different from the ones who have not. With the first group, we move faster, we build better, and they actually get to lead their team into AI. With the second group, every decision is a bottleneck because they cannot judge what they are looking at. You can still bring in vendors like us afterwards. We will do it the right way. But you will be a better leader for having tried it yourself, and your team will be a better team for having someone inside who already gets it. No fear. No dependence. Just better work, on both sides. If you run an independent hotel and you have not opened an AI yet, reply "start" and I’ll send you the exact first prompt I use to get owners from zero to a working agent in under an hour.
-
Piers Thackray a partagé ceciMy hotel's AI refused an order from me today. And it was right to. A guest paid a €207 deposit. Our payment link was broken on May 12, so I assumed nothing came through. I told the AI to email him and ask him to pay again. It stopped me: "I need to stop you here. This is a live, verified payment. Sending a 'try again' email will double-charge him €207." It had already pulled the charge from our live Stripe account. Status: succeeded. Paid. Not refunded. The guest had retried on May 15. The money was sitting in our account. I just hadn't seen it. So I pushed back. I told it the payment did not go through. It held its ground. It pulled the receipt again. It walked me through the dates. It told me to open the Stripe link and check for myself. Then it said one line I keep thinking about: "I won't send a 'retry' email while a matching live charge exists." Here is how I now think about AI at work: AI does what you expect. That's good. AI does what you don't expect. That's bad. AI does what you don't expect, turns out to be right, and holds its ground when you push back. That's the one you want. Most people are building AI yes-men. Tools that agree and execute. I'd rather have an AI employee. One that catches my mistake before it reaches a guest. One that over-delivers on the job I hired it for. Today it saved a guest from a €207 double charge. It saved me an angry email and a refund too. We push our R&D AIs hard. Some days it goes wrong. Today it went very right. If you're putting AI into your business, ask one question: does it just obey, or does it think?
-
Piers Thackray a partagé ceciWe rebuilt the same hotel AI 5 times in 2 months. The work we skipped to "save time" would have taken 2 days. Third property we ever put AI into. No real process. No clear sequence of decisions. Back then, we did not even have a definition of "process." So we built. Then we found a gap. Changed the process. Which meant rebuilding the AI. We did that five times. For two months. To fix something we could have mapped in a weekend. Here is what it taught me. A process = a series of decisions. Each decision = a set of parameters you weigh up before you choose. Simple decision = few parameters. Complicated decision = lots of parameters. Short process = few decisions. Long process = lots of decisions. Building the automation is the easy part. You connect a few tools and write some system prompts. Defining the process is where the real work happens. You either do that work up front on paper, Or you do it later by rebuilding broken automations in production. The rule of thumb we use now: 1 hour mapping a clear process = 30 hours saved building it. So before you "automate operations," do not call the whole thing complicated. Find the 2 or 3 decisions that are actually hard. Map those properly. Then build once.
-
Piers Thackray a partagé ceci"I don't want to commit to AI because I'm afraid of vendor lock-in." A hotel owner told me this last week. It is a reasonable fear. If all the knowledge sits with a vendor, you are exposed. That is why I do not treat AI like a SaaS subscription. When you do this properly: - Your team understands how the system works - Your hotel owns the accounts, data, and workflows - If you ever want to switch providers, you can In that setup, "lock-in" is mostly a paperwork problem, not an existential risk. The second mistake is thinking AI is just another tool in the stack. AI is an infrastructure layer. It touches how guests email you, how you track bookings, how you handle events, how you follow up on payments. Same as electricity. Not the coffee machine. So you really have three options: ✘ Ignore it and slowly fall behind. ○ Bet big and hope you picked the right vendor. ✓ Or make careful, low-risk moves: start with one contained workflow, measure hours saved, keep what works, discard what does not. That third path is the only way I would run a hotel. Follow for more content.
-
Piers Thackray a partagé ceciA GM asked me last week if Claude could manage MLOS for his hotel. Good question. So I built an agent on 3 years of our PMS data. Then ran it. MLOS is the quiet revenue lever in our hotel. Too tight and we turn away short-stay guests. Too loose and a 1-night booking blocks a 3-night window I would have sold for triple. Setting MLOS by hand across every room, every date, every season? Nobody has time for that. And getting another tool in my stack is not a priority. So I wired it up. The prompt: "Please review seasonal MLOS." Here's what happened. Claude pulled the next 90 days of availability out of my PMS. Ran my rules. Found 67 changes across 19 date intervals. Showed me each one with its reason. Asked for approval. A few examples from the run: ○ MAX1 Deluxe Suite, May 9-14: MLOS 1 → 2 (rolling back yesterday's last-minute drop) ○ MAX2/3 Sea-View King, May 26: MLOS 3 → 1 (1-night gap between two bookings, let's sell it) ○ MAX5 Standard w/ Terrace, May 30 to Jun 4: MLOS 3 → 1 (over-restricted) It also flagged what NOT to change. Where MAX5 could already absorb demand on a tricky date, it kept the premium room at baseline. Where a date was sold out, it skipped the write entirely. I reviewed. I approved. It pushed in one API call. 67 changes. Maybe 2 hours of work if I did it by hand on a Sunday morning. Done in the time it took me to read the proposal. Now here's the lesson, because it matters more than the tool. One prompt was enough because the infrastructure was already there. The PMS API was connected. Claude could read and write to my systems. The data flow was in place. The rules were already written down. So adding MLOS was simple. Reuse the rules. Wire the prompt. Ship. The hard part is always the first one. Once the foundation exists, you stack new automations on top fast. Especially the boring ones. Clear input. Clear output. Clear rules. Other things still take real work. Guest communication needs a lot of data and a lot of tuning. But the dry operational stuff? One hour each, once the rails are laid. This is what most hotel owners miss. The first automation feels slow. The tenth one feels free. If you're sitting on a task you've been putting off because it's tedious, don't go find a vendor. Build the infrastructure once. On your rules. Follow if you want the next one. I'm wiring up alerts next.
-
Piers Thackray a partagé ceci“We don’t really have to lift a finger anymore.” Big thanks to Christian Ullmark and the team at Hotel Slottsvillan in Sweden for recording this in‑the‑garden testimonial and for letting me be part of what they’re building. This project has been one of my favorites so far. Before: - 100+ wine‑tasting emails coming in - Manual tracking in sheets - Manually checking payments, chasing changes, updating bookings - A lot of receptionist / admin time going into low‑value back‑and‑forth What we ended up with: - Requests come in - The AI replies to guests, processes payments, and sends confirmations - It can book rooms and handle date changes when guests need to move their stay - Next up: booking restaurant tables and placing guests to maximize table usage and profit Net effect: - Guests get faster, more consistent answers - Christian’s team spends more time on marketing and taking care of guests, less on admin - The back office runs a bit calmer I’m really proud of what Christian and his team have done with this so far, and excited about where we can take it next.
-
Piers Thackray a partagé ceciIf you run Cloudbeds, Mews, OPERA Cloud, Clock PMS+, or Guesty, Claude can now actually run parts of your hotel, not just chat about it. I've built small tools that give Claude a safe "key" into each PMS. Translation: your AI can now work inside your system instead of just talking around it. In practice, that means Claude can: ○ Pull tomorrow's arrivals before your coffee. ○ Spot rate drift across the next 30 days and suggest fixes. ○ Reply to email inquiries with the right rates and rooms. ○ Build today's cleaning plan from live bookings. ○ File receipts and ship the weekly revenue report. All inside your PMS. All on its own. Most PMS APIs are "open" on paper but painful in reality. Custom auth, weird pagination, partial schemas. They break normal AI tools. So I built one CLI per PMS that hides all that and gives Claude a clean, hotel-operator-friendly surface to work with. Every CLI here has an MCP server built in, so Claude can use it the moment you install it. And your team can use it too, just by typing one command. Today the list covers: 1. Cloudbeds 2. Mews 3. Oracle OPERA Cloud / OHIP 4. Clock PMS+ 5. Guesty 6. 15+ more PMSs in the pipeline (coming soon) Well over 700 endpoints across the five. All open source. One-line install. If your PMS is on that list, comment "Hotel CLIs" or your PMS name and connect with me. I'll DM you the repo so your team can put Claude on top of your PMS this week. Don't see your PMS? Comment it below and connect with me. I'm working toward "every hotel PMS, no exceptions" and will build yours next and send it back to you.
-
Piers Thackray a partagé ceciOur hotel rating moved from 9.1 to 9.5. The reason surprised me. For the past year, I've been building AI automations inside our boutique hotel. The plan was simple. Take boring work off our plates so the team could spend more time on the guest. What I didn't expect was for the rating to move this much. So I went looking for why. Three things stood out. 1. The team is lighter on its feet. I asked the team how they felt about the work since we started rolling out AI. Most of them said the same thing. The copy-paste work is gone. The reports. The data shuffling. The same email written for the 200th time. They start the day less drained, less stressed, less overwhelmed. I can't prove this with a dashboard. But guests feel it. A team that isn't burned out shows up differently at the front desk. That alone moves a rating. 2. We deliver services we never used to deliver. I dove into our email analytics to see what had actually changed. Before, we sent zero proactive trip-planning emails. Now those emails go out to every VIP booking before arrival, with offers to help on lunches, dinners, beach clubs, drivers, tours, museums, the private chef. The team keeps the overview. The AI does the prep work. Guests arrive feeling like the hotel was already thinking about their week. That used to be reserved for guests who picked up the phone and asked. Now it's the default. 3. We spend money on what guests actually feel. I dove into the finances to look at how our spending had shifted. Two things had moved. Cost of repetitive processes was down. Investment in amenities was up. The hours saved on admin freed up budget for things that show up in reviews. Like paying a premium for a fixing team always on call. Small hotels rarely have a full-time handyman. We effectively do now, because the math finally works. A broken shower fixed in 30 minutes is a five-star review. A broken shower fixed in two days is not. Three layers, in this order: ○ Free up the team. ○ Deliver the services you couldn't deliver before. ○ Reallocate the savings to what guests actually feel. I didn't set out to move the rating with AI. The rating moved because the team finally had room to do the work they always wanted to do. If you run a small hotel and you think AI is mostly about answering emails faster, you're missing the bigger story.
-
Piers Thackray a partagé ceciMy hotel runs on 21 AI agents. None of them existed a year ago. Some run on a schedule. Some I trigger by name. A few wait for emails or bookings to come in. Here's the full list, grouped by what they actually do. DAILY OPS (1) ▪ morning-routine: runs every agent below that needs to fire before I open my laptop. EMAIL & COMMS (5) ▪ email-triage: scans 3 inboxes, classifies, prioritises. ▪ email-processor: drafts replies, logs them to a sheet. Auto-runs every 15 min. ▪ email-draft: writes one-off emails in my voice. ▪ booking-inquiry: warm, personal replies to inquiry emails. ▪ arrival-time-check: weekly chase for missing arrival times. GUEST EXPERIENCE (2) ▪ concierge: restaurants, tours, transfers. Vetted contacts only. ▪ pre-arrival-outreach: welcome messages for premium bookings. REVENUE & PRICING (3) ▪ dynamic-pricing: proposes rate and MLOS changes for the next 30 days. ▪ room-upgrade: spots upgrade chances. Executes after I approve. ▪ agent-notification: tells rental agents when rates or availability move. HOUSEKEEPING & F&B (2) ▪ cleaning-schedule: daily plan built from Cloudbeds bookings. ▪ fb-shopping-list: grocery list from guest counts plus inventory. FINANCE & REPORTING (3) ▪ finance-filing: reads receipts, files to Drive, logs to a sheet. ▪ villa-analytics: pulls and analyses booking and revenue data. ▪ marketing-report: 2-page weekly doc on traffic, ads, and occupancy. SALES (1) ▪ crm-pipeline: tracks every inquiry from first message to post-stay. Drafts follow-ups for stale leads. BUILD & RESEARCH (4) ▪ Explore: fast search across my code and files. ▪ Plan: designs the steps before any non-trivial build. ▪ general-purpose: fallback when nothing specialised fits. ▪ web-research: cheap online research for suppliers and contacts. The whole stack cost me less to build than one front-desk month salary. Comment "hotel agents" and I'll send you the entire repo. All 21, ready to plug into your Claude.
Piers Thackray a répondu à un commentaire
1 j.
Michaël Desplaces Sent!
Piers Thackray a répondu à un commentaire
4 j.
Damian LAZAR Please connect so I can send you the advisor AI via message.
-
Piers Thackray a aimé ceciSo grateful for the HSMAI BC community as a remote employee!Piers Thackray a aimé ceciA great time was had last night! Big thanks to Chance McKee for sharing his wisdom and our sponsors Vancouver Fraser Port Authority and FortisBC for the amazing food and drinks at Hilton Vancouver Downtown. Expect more photos from the talented Oleg Donner of New Globe Productions in the coming days, but we just had to share a teaser. Thanks for coming out Chantal Jones, Brandi Lasanen, Natalie Nicholls, Natalie Kim-Kamal, Luis Riestra, Pat Miller, Anthony Woo, CRME, Lien Chang, Ramy Fouda, M.S., CRME, CHA, Kristian Miller ~ CHA, CHGM, Anna Milenkova, CRME and the Hallmark Hotels team, Tina Kyriakopoulos and Rie Kobori! #HSMAIEvents #HSMAIBC #TourismBC #BCCruiseIndustry #BCHospitality
-
Piers Thackray a aimé ceciPiers Thackray a aimé ceciThe boutique revolution is rewriting hospitality: bespoke, sustainable, tech-savvy, and unapologetically personal. Big chains are scrambling to keep up. The future of travel? Authenticity, not cookie-cutters. #HospitalityTrends #BoutiqueHotels #SustainableTravel #TechInHospitality #BleisureTravel Practice Hospitality ManagementThe Next Big Thing in Hospitality? Thinking Small.The Next Big Thing in Hospitality? Thinking Small.
-
Piers Thackray a aimé ceciPiers Thackray a aimé ceci🇮🇹 Visita a las oficinas de Milán de The Palace Company Tuve la oportunidad de visitar el corporativo de Milán, una experiencia muy enriquecedora que me permitió aprender de cerca del equipo de Europa y fortalecer la colaboración en proyectos clave como Global Sales y HubSpot. Esto un paso importante en la transición para seguir impulsando nuestra operación global. En lo personal y profesional, me llevo grandes aprendizajes, una visión más global del negocio y la motivación para seguir impulsando estas tecnologías. Gracias igual al equipo de Mérida por su soporte y rápida respuesta a las necesidades del equipo en Milán, contribuyendo a una transición más ágil y eficiente.
-
Piers Thackray a aimé ceciPiers Thackray a aimé ceciAt BAHAUS Interior Design, we believe that a home should be more than beautiful — it should be iconic. That's why we're proud to announce our exclusive collaboration with Versace Ceramics, bringing the world's most distinguished design house into the spaces where life is lived. Floor-to-ceiling light. Marble that breathes. The Medusa in every detail. This is branded residences — reimagined. #HotelEquities #Hospitalitylndustry #AssetManagement #HotelInvestment #HospitalityLeadership #HotelDevelopment #RealEstate #HospitalityTrends
-
Piers Thackray a aimé ceciMy choice for the next stay in Prague! Looking forward to exploring the Golden City and then be pampered at the Fairmont Golden PraguePiers Thackray a aimé ceciWe’re keeping the momentum going! Fairmont Golden Prague has officially opened its doors, marking an extraordinary new chapter for Fairmont Hotels & Resorts in the heart of Prague 🦅 This remarkable hotel is far more than a new addition to our portfolio — it’s a celebration of heritage and reinvention. Originally built in the 1970s, this historic building in Prague’s UNESCO World Heritage-listed Old Town has been thoughtfully reimagined as a Fairmont. With its unparalleled location, landmark architecture, and a rich history reflected in its detailed design and curated art collections, it’s the perfect representation of our heritage properties in Europe — a true work of art, inside and out. Guests are invited to experience six unique culinary destinations, starting with Zlata Praha and Golden Eye, offering breathtaking rooftop views of Prague, CooCoo’s Nest, which is connected to the hotel’s own art gallery, and Greenhouse — an upcoming favorite for locals! In addition to the 320 rooms, suites and residences, Fairmont Golden Prague also unveils a 1,400 sqm Fairmont Spa & Wellness, a signature Fairmont Gold experience and a striking 600-square meter column-free ballroom set to host grand celebrations. 💻 📱 Book your stay on our new Fairmont website - https://lnkd.in/eTBNTrT5 #FairmontHotels #FairmontGoldenPrague #MakeSpecialHappen #Hospitality #Accor #CzechRepublic
Expérience et formation
-
TH1.ai
Voir toute l’expérience de Piers
Découvrez son poste, son ancienneté et plus encore.
ou
Licences et certifications
Expériences de bénévolat
-
Volunteer
World Peace Forum
- aujourd’hui 10 ans
I help organising the World Peace Forum and the Luxembourg Peace Prize event.
-
Humanitarian Worker
Globalteer
- 1 mois
Enfants
I took care of orphans in Medellin, Colombia ranging from 3 years old to 16 years old, I also taught them English and break dance.
Langues
-
English
Bilingue ou langue natale
-
French
Bilingue ou langue natale
-
Spanich
Compétence professionnelle limitée
