To automate pre-arrival and post-stay emails, use your property management system's scheduled messages, or an email tool linked to your bookings, triggered by check-in and check-out dates. Use AI to write the templates once from a property fact sheet and to draft replies when guests answer, rather than generating every email from scratch.
You'll come away with a six-message timeline, templates for each message, rules for sending different versions to different guests, and a safe way to let AI help with replies. The caveat worth knowing now: a thank-you email and a "come back for 15% off" email are treated differently under data-protection and marketing rules, so the post-stay half needs a little more care than the pre-arrival half.
A six-message timeline that covers most stays
Most properties need six messages, and the booking confirmation is usually already handled by the booking engine. The timings below suit hotels, guest houses, holiday lets and campsites; adjust the days to your typical booking window.
| Message | When | What it does | Personalised with | AI's job |
|---|---|---|---|---|
| 1. Confirmation | On booking | Confirms dates, price, cancellation terms | Booking details | None; your booking engine sends it |
| 2. Planning your stay | 7 days before | Directions, parking, what to bring, asks arrival time, offers extras | Room or pitch type, nights, party | Writes the template and variants |
| 3. See you tomorrow | 1 day before | Check-in time and method, key or door code, late-arrival plan | Arrival time given, check-in method | Writes the template |
| 4. Settled in? | Evening of arrival or next morning | Invites problems early, while you can fix them | First name | Drafts replies to what comes back |
| 5. Thank you | 1 day after check-out | Thanks, asks for a review, invites private feedback | Stay dates, room | Writes the template; summarises feedback |
| 6. Come back | Weeks later, or before the same season next year | Offer for a return stay | Previous stay type | Writes variants by guest type |
Message 4 earns its place. A guest who tells you on the first evening that the shower runs cold gets it fixed and leaves happy; one who discovers it and says nothing tells everyone in a review. Send it by text or WhatsApp if your system supports it, since guests rarely check email mid-stay. If it's WhatsApp, check what your system passes on in costs: property systems usually send through the WhatsApp Business Platform, which charges per message at rates that vary by the guest's country, and from 1 October 2026 utility templates sent inside the 24-hour customer-service window are charged too, while free-form replies in that window become chargeable after the first 1,000 a month per number. For a small property that's rarely much money, but it's worth knowing before you switch every message to WhatsApp.
Where the automation should live
If you use a property management system (PMS), start there. Most include scheduled guest messages and some now include AI:
- Cloudbeds describes automations driven by reservation data, with conditions such as room type, booking source, length of stay and guest history, and an AI assistant for guest conversations trained on your own property information.
- Little Hotelier, aimed at small properties, offers customisable pre-arrival and post-stay email templates it can send automatically.
- Hostaway, common with holiday-let managers, has message automations plus AI Replies that suggest responses in the inbox, with a tone setting, a knowledge base of property facts, and an AI Auto-Reply option for sending answers automatically.
If you don't have a PMS, for example a campsite taking bookings on a simple calendar, you can build the same timeline with an email tool and an automation platform: a new row in your booking sheet triggers Zapier or Make, which schedules emails based on the arrival date. It works, but you have to maintain it, and it breaks quietly when a booking is moved. Where a PMS exists, use it; where it doesn't, keep the do-it-yourself version to messages 2, 3 and 5.
For a side-by-side look at messaging tools for small properties, see hotel AI guest messaging tools compared.
Write the templates with AI, once, from a fact sheet
The quality of every automated email depends on the facts behind it. Start with a fact sheet for the property (or one per property if you manage several):
PROPERTY FACT SHEET
Name / type: [e.g. 12-room boutique hotel in a converted townhouse]
Check-in: [time], method: [reception until 10pm / key safe / app]
Late arrival: [what happens after reception closes]
Check-out: [time]; late check-out: [available? price?]
Parking: [where, cost, height limits, EV charging]
Getting here: [nearest station, taxi time, last 200 metres that confuse people]
Breakfast: [times, price if extra, dietary options, how to pre-order]
Extras you sell: [e.g. bottle of wine $30, dog bed, cot, early check-in $20]
House rules that matter: [quiet hours, no smoking, pets policy]
Tone: [e.g. relaxed, warm, short sentences, British spelling]
Never promise: [e.g. specific room numbers, early check-in, free upgrades]Then ask a chat assistant to write each message:
Using the fact sheet above, write message 2 ("planning your stay"),
sent 7 days before arrival. Use these merge fields exactly as written:
{first_name}, {arrival_date}, {nights}, {room_type}.
Include: how to get here, parking, one question asking arrival time,
and a short list of extras with prices. Under 180 words.
Write a second version for guests arriving with a dog.
Use only facts from the sheet; if something is missing, write [CHECK].A realistic first draft for an eight-room guest house (sample output, illustrative):
Hi {first_name}, we're looking forward to welcoming you on {arrival_date} for {nights} nights in the {room_type}. We're just a five-minute walk from the station, or a short taxi ride. Free parking is available on site [CHECK: parking cost]. Could you let us know roughly what time you'll arrive? If you'd like to add anything to your stay: a bottle of wine ($30), early check-in ($20), or a dog bed for your furry friend.
Three fixes. "A five-minute walk from the station" isn't on the fact sheet, which says ten minutes by taxi; the assistant has smoothed the facts into something friendlier and wrong. "Free parking" sits next to a [CHECK] marker for the parking cost, so the draft contradicts itself: delete "Free" until you've filled the marker. And the dog-bed line belongs only in the dog-owners variant, not the standard one, or every guest gets offered a dog bed. The merge fields are right, but compare them character by character with your system's names anyway; {firstname} and {first_name} look the same at a glance.
The "[CHECK]" instruction is the useful part. Instead of inventing a parking price, the assistant leaves a marker you'll spot. Before loading any template, read it against the fact sheet line by line, and swap the merge-field names for the exact ones your system uses. For how to present extras without sounding like an airline, upselling upgrades and late check-out with AI goes further.
Send different messages to different guests
One template for everyone is where automated emails start to feel automated. A handful of conditions covers most differences:
- Booking source. Direct guests may already have your directions from the booking page; online travel agency guests often haven't read anything. Give the second group a fuller message 2.
- Length of stay. A one-night guest doesn't need the list of walks; a seven-night guest does.
- Party. Dogs, cots, accessibility needs flagged at booking: each gets its own paragraph.
- Arrival time. If a guest has said they'll arrive after reception closes, message 3 should lead with the late-arrival instructions, not bury them. The standard version opens "We're looking forward to seeing you tomorrow. Check-in is from 3pm at reception", and the late-arrival code sits in paragraph three. The late-arrival version opens "You mentioned arriving around 11pm, after reception closes at 10, so here's how to let yourself in: the key safe is to the left of the main door, code {door_code}, and your key is in envelope {room_number}." Same facts, reordered for the guest who'll be reading them on a phone in the dark.
- Repeat guests. Swap "how to find us" for "welcome back" and mention anything that's changed since their last stay.
At a campsite the conditions look different but the idea is the same: tent pitches get the ground and pegs advice, caravan and motorhome pitches get the electric hook-up and waste-point details, and anyone arriving after the barrier closes gets the gate code procedure first. At a boutique hotel, a guest who booked a dinner package gets restaurant timings in message 3.
Let AI draft the replies, not send them, at first
Good pre-arrival emails get replies: "we'll arrive around 9pm", "can we check in early?", "is there a fridge in the room?". This is where AI saves real time. In a PMS with AI replies, or in Gmail or Outlook with their built-in assistants, the AI reads the guest's message and drafts an answer from your property facts.
A typical early draft shows why (sample output, illustrative). The guest writes: "Hi, we're hoping to get there around 11am. Any chance of checking in early? Also, is there a fridge in the room?" The AI drafts: "Hi [name], checking in at 11am is no problem at all! Every room has a mini-fridge, so you can keep drinks cold." Both sentences are wrong for this property. Early check-in depends on the room being ready and costs $20 on the fact sheet, so it can't be promised; and only the two superior rooms have fridges, a fact nobody had put in the knowledge base. The corrected reply: "Hi [name], we'll do our best to have your room ready early. Early check-in is $20 when it's available, and we'll confirm on the morning of your arrival. Your room doesn't have a fridge, but we're happy to keep anything cold in ours." That one exchange produces two entries in the correction log: a rule ("never confirm early check-in") and a fact ("fridges: superior rooms only; others can use the house fridge").
Run it in draft mode for the first four weeks. You read each draft, fix it and send. Keep a note of every correction: each one is either a missing fact (add it to the knowledge base) or a rule (add it to the guardrails, such as "never confirm early check-in; say we'll try and confirm on the day"). After a month, you'll know which reply types the AI gets right every time, usually arrival times, parking and breakfast hours, and you can allow automatic replies for those alone. Anything involving money, complaints, accessibility or a change to the booking stays with a person.
Service emails and marketing emails follow different rules
Messages 2 to 5 are about a stay the guest has booked, and guests expect them. Message 6 is marketing, and under data-protection law such as the GDPR and related marketing rules, promotional emails generally need a lawful basis such as consent. The practical steps:
- Add a clear, unticked opt-in to your direct booking form and your check-in form: "Send me occasional offers for return stays."
- Send message 6 only to guests who opted in, and include an unsubscribe link in every one. Write a variant per guest type rather than one offer for all. For guests who stayed with a dog, for example: "You stayed with us last autumn with your dog, and the lane walks are at their best again in October. If you'd like to come back, the dog-friendly rooms for October are open for booking now, and returning guests get the second night at half price." The couple who stayed two nights midweek gets a midweek offer; the family gets the school-holiday dates. The AI is good at producing these variants from one brief, as long as the offer terms come from you.
- Treat online travel agency guests carefully. Booking.com relays messages through an alias address ending in @guest.booking.com and doesn't share the guest's own address as standard; check your agreement before sending anything promotional through it. Asking for consent at check-in is the clean route.
- Keep review requests in message 5 honest: ask everyone, offer no incentive, and don't send only happy guests to public review sites. Review gating explains why filtering guests that way can get your listing penalised.
If you're unsure about consent wording for your situation, ask your data-protection adviser; the rules differ by where your guests live and how you collected their details.
Fourteen cottages, one manager: how it runs
Say a holiday-let manager looks after 14 cottages for different owners, with about 45 arrivals a month in season. At six messages per stay that's roughly 270 messages a month. Sent by hand at about four minutes each, including looking up the cottage details, that's around 18 hours a month, and it's the job that slips on changeover days.
Setup (about two days): one fact sheet per cottage, built from the owners' information packs and checked on the next visit. Templates for messages 2 to 6 written with AI, with variants for dogs and for stays under three nights. Automations set in the PMS, triggered by arrival and departure dates, with booking source as a condition.
Weeks 1-4: AI drafts replies; the manager checks each one. The corrections are mostly cottage-specific: which one has a step-free entrance, which key safe sticks in wet weather. Each goes into that cottage's fact sheet. After four weeks, one cottage's sheet has gained lines like these (illustrative): "Key safe: stiff when wet; push the dial in before turning." "Step-free: no; two steps at the front door, level access at the back via the garden gate." "Wi-fi: weak in the back bedroom; the router is in the kitchen cupboard, restart it there." "Nearest shop: the village shop closes at 5pm on Sundays; the next one is 15 minutes by car." Every one came from a guest question the AI couldn't answer, which is the best way to find the gaps in an owner's information pack.
After a month: sending is fully automatic; reply drafting takes around 20 minutes a day instead of several scattered hours. The manager also notices that guests who got the "settled in?" message report problems on night one rather than in the review, which is exactly what it's for.
How to check the automation is working
- Before launch: create test bookings for next week with your own email address, one for each guest type. Check every message arrives on the right day, at a sensible hour in your time zone, with every merge field filled.
- Weekly for the first month: skim sent messages for any "[CHECK]" markers or blank fields that slipped through, and any booking that was moved but still got messages for the old dates. The moved booking is the classic failure of a do-it-yourself set-up (illustrative): a guest moved her stay from the 12th to the 19th, the booking sheet was updated, but the automation had already scheduled messages for the 12th. She received "see you tomorrow", with the door code, a week early, and replied confused. The fix in Zapier or Make is to check the arrival date again at the moment of sending: a filter step that stops the message if the booking row's date no longer matches. In Zapier a filter step doesn't count towards your task allowance, so the check costs nothing to run.
- Monthly: track the share of guests who reply with an arrival time, take-up of each extra, the number of reviews, unsubscribes from message 6, and complaints that a message should have prevented.
- Each season: reread every template against the fact sheet. Parking prices, breakfast hours and door codes change, and an automated email carrying last year's code does more harm than no email.
Once the pre-arrival half runs itself, the post-stay half is where the business value is: return bookings from guests who already know you, without paying commission on them. For the review side, see getting more reviews with automated requests; for turning past guests into direct bookings, the direct-bookings tutorial in the reading list below picks up where this one stops.
Guest email automation questions
Should the AI write each email fresh for every guest?
Usually not. Fixed templates with merge fields, written once with AI and checked by you, are more reliable and easier to fix. Let AI personalise only where it has trustworthy booking data, such as the room type or the number of nights, and use it mainly to draft replies when guests answer. Emails generated from scratch for every booking can drift from your facts without anyone noticing.
How many pre-arrival emails are too many?
Two before arrival is enough for most stays: one about a week ahead with directions and choices, and one the day before with check-in details. Add a third only for long or complex trips, such as a group booking or a pitch that needs a vehicle registration in advance. Each extra message should carry information the guest needs, or reply rates fall and important details get skimmed.
Can I send post-stay emails to guests who booked through an online travel agency?
You can send service messages about the stay through the channel the platform provides. Booking.com, for example, relays email through an alias address ending in @guest.booking.com rather than sharing the guest's own address. For anything promotional, check your platform agreement and data-protection rules first, and ask guests at check-in whether they'd like to hear about offers directly.
What if a merge field is empty, like a missing first name?
Set a fallback in your system so the email says "Hello" or "Hello there" instead of "Hi ,". Most property management systems and email tools let you define a default value for each field. Test it by creating a booking with the name field left blank and sending the message to yourself before you switch the automation on.
Further reads
- How Small Hotels Use AI to Answer Guest Messages Around the Clock — Answering guest messages around the clock once emails go out.
- How Campsites and Glamping Sites Use AI to Answer Guests — Campsite-specific questions, from hook-ups to gate codes.
- How Independent Hotels Win More Direct Bookings With AI — Turn post-stay emails into direct repeat bookings.
- How to Reply to Hotel Reviews With AI Without Sounding Canned — What to do with the reviews your post-stay email brings in.
- How to Segment Your Email List With AI, Even a Small One — Split past guests into groups worth different offers.
- When Should an AI Chatbot Hand Over to a Human? — Decide which guest replies must reach a person.
- How to Write B&B Room Descriptions and Guest Guides With AI — Room fact cards, prompts for each listing, the words that lead to refund disputes, and a quick way to turn a walk round the house into a guest guide.
- How B&B Owners Can Automate Airbnb and Booking.com Messages — A message-by-message plan for automating B&B guest messages on Airbnb and Booking.com with built-in tools, AI-written templates and a channel manager.
- Where Should a Small Hotel Start With AI? — A scoring grid for picking a small hotel's first AI job, starting points by property size, and a 30-day plan for an 18-room hotel.
- Can a Hotel Chatbot Take Bookings or Only Answer Questions? — The three things a hotel chatbot can do with a booking enquiry, what each needs from your booking engine, and which suits a small property.
- After an AI Consultation: Turn the Advice Into a 30-Day Plan — A one-page AI action plan template, a filled-in 30-day plan for a boutique hotel, the owner's weekly check and how to judge the result at day 30.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Cloudbeds product pages on guest messaging and automations; Little Hotelier product pages on pre- and post-stay emails; Hostaway support pages on AI Replies and message automations; Booking.com partner help on guest alias email addresses, all checked September 2026. Volumes and times in the worked example are illustrative.