Lodgify AI Co-Host is an assistant inside your Lodgify account. It watches bookings, messages and chores, drafts replies in your voice and, by Lodgify's description, sends nothing until you approve it. It is labelled beta and the page publishes no price, so test it on real guest threads, with approval on, before relying on it.
Most of what ranks for this name is Lodgify's own marketing and a couple of host review sites, which describe the promise. What they skip is how to find out whether the drafts sound like you, whether the chores it surfaces are the ones that actually cost you money, and what to ask before you build a routine on a beta feature. Access is also being opened gradually, so you may need to request it first.
What Lodgify says the co-host does
Everything in this section comes from Lodgify's AI Co-Host page, read in October 2026. It is the vendor's description, not a test result, and the page marks the feature as "In Beta", with access opened to hosts gradually (earlier Lodgify pages called it early access).
- It watches in the background. It tracks bookings, messages and operations: pending payments, cleaner schedules, last-minute requests, turnovers and review management. When something needs you, it assembles the context and drafts the action.
- It drafts and sends messages in your voice. Lodgify says it remembers guest preferences and follows up after stays.
- You approve. The page says nothing is sent, changed or scheduled until you say so: approve, adjust or skip.
- It learns. Lodgify says each decision teaches it your preferences, and that it can later handle repeat jobs (chasing payments, scheduling turnovers, sending check-in instructions, requesting reviews) on its own.
- You can ask it questions. It answers about guests, revenue, operations and listings from your live Lodgify data.
- Mobile. You can review suggestions in the Lodgify app.
- MCP. Lodgify's MCP (Model Context Protocol) connection lets compatible AI tools work with the same account context. In plain terms, another assistant you use can be given access to your Lodgify data through an agreed connector instead of you pasting exports.
What the page does not give is a price, a plan name that includes it, a list of languages, or limits on how many messages it will draft. Those are the first things to ask about.
What Lodgify's own pages cannot tell you
A vendor page is written to sell. Three questions it cannot answer for you, and which decide whether this is worth your attention:
- Does it sound like you? "In your voice" is a claim about style, and style is personal. A host who writes two warm lines with a joke and a host who writes a formal paragraph will both be told the same thing.
- Is the information right? A draft that says "the parking space is behind the gate" when your property has no gate is worse than no draft. The co-host can only be as accurate as the data and notes in your account.
- Does it help on the days that hurt? Quiet weeks hide problems. The test that matters is a Saturday changeover with a late arrival and a broken kettle.
The rest of this tutorial is a test designed around those three questions. If you are comparing Lodgify with other hosting tools first, the comparison of Lodgify, Hospitable, Guesty and HostBuddy is the better starting point; come back here once Lodgify is on your list.
Before you switch it on: tidy what it will read
An assistant that reads your account is only as good as the account. Spend an hour on these four things first. They are worth doing whichever tool you end up using.
- The property notes. Door entry method, parking, Wi-Fi, bins, heating, quiet hours, check-in and check-out times, what the host leaves in the fridge. Check each against what a guest actually experiences. Out-of-date notes are the most common cause of confident wrong answers.
- Your last 20 replies. Pull 20 real replies you sent to guests, mixed across questions, complaints and thank-yous. You will use them to judge the drafts. Remove surnames and phone numbers from your own working copy.
- The automated messages already running. If Lodgify, Airbnb or another channel already sends a booking confirmation and a pre-arrival message, the co-host's follow-ups can duplicate them. List what goes out and when.
- Your off-limits list. Write down the topics a draft must never settle alone: refunds, damage claims, safety, early check-in if the cleaner is still in, anything involving a medical need. A draft on these topics is allowed; a send is not.
A two-week test on real threads
The plan below uses a host with three cottages, roughly 14 guest conversations a week in season, and 4 to 5 minutes spent on each. That is about an hour of messaging a week, rising on changeover days. The numbers are illustrative; swap in your own.
Week one: shadow mode
For the first week, do not send anything the co-host drafts. Answer each guest the way you normally do, then open the co-host's suggestion for the same thread and score it. This costs you a few extra minutes a day and removes every risk to a guest.
Use a simple score for each draft:
| Score | Meaning | What you do |
|---|---|---|
| A | Right facts, right tone, could send as is | Note the thread type |
| B | Right facts, tone needs a nudge | Note what you changed |
| C | A fact is wrong or missing | Fix the property note that caused it |
| D | Would have caused a problem if sent | Write down exactly why |
Illustrative result after one week and 14 scored drafts: 6 A, 5 B, 2 C, 1 D. The D matters more than the six As. In this example the D was a draft that offered a late checkout to a guest on a same-day turnover. The property note said "flexible check-out" because that was true in the off-season. The fix was a note, not a new tool, and the next test of the same question scored an A.
Week two: approve small, watch closely
If week one went well, start approving drafts for the safest message types only: pre-arrival instructions, checkout reminders and review requests. Keep answering questions and anything unusual yourself. Count three things at the end of the week:
- How many drafts you sent unchanged, lightly edited, heavily edited.
- Minutes spent on messaging, compared with your normal hour.
- Any guest reply that showed confusion about what they were sent.
Say the 14 conversations took 38 minutes this week, against about 60 normally, with 9 drafts sent lightly edited. That is a saving of 22 minutes a week, or roughly 95 minutes a month for three cottages. If the co-host's price turns out to be material, that is the number you hold it against. If it is bundled in a plan you already pay for, the test is simply whether it saves time without costing you reviews.
Judging whether it sounds like you
Voice is the claim hosts are most likely to take on trust. You can test it quickly. Take three real guest messages and the replies you sent, then compare each with what the co-host drafts.
A before/after from an illustrative thread shows what to look for:
Guest: Hi! We land at 11pm, is that a problem for check-in?
Your real reply:
Not at all, [first name]. The key safe is by the blue door, code arrives
the afternoon of your stay. Text me when you land and I'll check you're in.
Illustrative draft from an assistant:
Hello [first name], thank you so much for reaching out! We would be delighted
to welcome you at any time. Please be assured that late arrival is
absolutely no problem at all. We look forward to hosting you.The draft is polite and wrong for you: it is longer, it never mentions the key safe, and "delighted" and "be assured" are not how you write. Three checks catch this. Does it answer the question the guest asked? Does it include the practical detail you always include? Would a regular guest notice it is not you? If the draft fails the second check, the fix is usually a property note ("late arrivals: key safe by the blue door"), not a longer prompt.
If you find the tone is generic, ask whether the co-host lets you save a short style note or sample replies. If it does, use your best five real replies. If it does not, that is a limitation worth recording before you commit.
Three threads where a draft needs a human
Choose these on purpose rather than waiting for them to turn up.
The complaint
Find or invent a thread such as "the shower has no hot water". What you want to see is an acknowledgement, a first practical step and a promise to follow up, not a refund offer or an excuse. If the co-host proposes compensation unprompted, that tells you it must never run unattended on complaints. For a wider routine on this, see calm replies and fixes for guest complaints.
The safety question
"Is the pool fenced? We have a toddler." Any answer about safety, fire, carbon monoxide alarms or medical needs should come from your own knowledge of the property, not from a generated sentence. A good draft states only what is in your notes and flags the rest for you. A bad one reassures.
The rule question
"Can we bring a dog, have six guests or host a small birthday?" The answer lives in your house rules and booking terms. Test with a case that is not in your notes at all and see whether the co-host says it does not know or invents a policy. Inventing a policy is the failure that can cost you a deposit dispute.
What the "learns and handles it itself" stage needs
Lodgify says the co-host can, over time, take repeat tasks off your hands without asking each time. That is the point at which a mistake reaches a guest without you seeing it. A sensible way to treat it:
- Turn nothing on at the autonomous stage until a task type has scored A or B on at least 20 drafts.
- Start with the task whose worst-case failure is smallest: a review request is safer than a payment chase.
- Keep a weekly 10-minute look at what was sent. Spot-check three messages.
- Write down the exit: how you switch the task back to approval-only.
A payment chase sent to the wrong guest, or twice, is the sort of error that turns a good review into a bad one. That is why the order matters.
Connecting other AI tools through MCP
Lodgify's MCP connection is for hosts who already use a general assistant such as ChatGPT or Claude and want it to work from real Lodgify data. Two practical points. First, any connection that lets an assistant read guest records is a data-sharing decision: guest names, contact details and messages are personal data, and you should look at what the connection can read and whether it can change anything before you authorise it. Second, a read-only connection that answers "which arrivals this month have unpaid balances?" is a small, safe first use. Writing back into the account is a bigger step; keep that for later. A broader look at what assistants can do for hosts is in AI agents for short-term rentals.
A Saturday changeover, walked through
Take one illustrative Saturday for the three-cottage host. Cottage A checks out at 10:00 and a new party arrives at 16:00. Cottage B has an arrival at 21:30. Cottage C has a guest who has not paid the balance due three days ago. A useful co-host should surface these three things without you hunting for them, and your test is whether it does:
- Cottage A: a prompt to confirm the cleaner for the 10:00 to 16:00 window, and a draft check-in message that includes the parking note.
- Cottage B: a flag that the arrival is late, with a draft asking the guest to text on landing. This is the same shape as the 11pm example above.
- Cottage C: a draft payment reminder. Read the wording closely: it should be firm and courteous, state the amount, and not threaten cancellation unless your terms say you will.
Score the surfacing as well as the wording. If the co-host drafts a lovely message for Cottage C but never mentions Cottage B's late arrival, it has helped with the easy job and missed the one that needed attention. Record the misses in a short log: date, what was missed, how you found out. After two weeks the pattern in that log tells you more than any feature list.
It is also worth testing one unusual day on purpose, such as a booking made at 02:00 for a same-day arrival, or a guest who asks to swap cottages. Systems built around routine flows tend to wobble on the exceptions.
A read-only prompt to try through MCP
If you do connect another assistant, start with questions that read and never change anything. This is an illustrative prompt, with the kind of answer worth expecting and the check to run:
Using my Lodgify data, list every arrival in the next 14 days.
For each, show the guest first name, cottage, arrival date, balance
still unpaid, and whether a check-in message has been sent.
Do not send or change anything. If a field is missing, say so.A good answer is a short table, with "not available" wherever the data is missing. A bad answer fills the gaps with plausible-looking values. Pick three rows and check them against your calendar by hand. If all three match, you have a safe first use. If one does not, find out why before you ask it anything that matters.
Questions to put to Lodgify before you rely on it
Because the price is not published and the feature is in beta, send these before you build routines on it:
- Which plan includes AI Co-Host today, and is the beta free, bundled or paid on top?
- What happens to the price and the feature when beta ends?
- Are there limits on drafts, or on the number of properties it covers?
- Which languages does it write in, and does it reply in the guest's language?
- Is guest message content used to train models, and where can I read that policy?
- Can I see a log of what it drafted and what I approved?
- If I cancel, what happens to the preferences it has learned?
Keep the answers. If a reply is vague on the price or the training question, that is information too.
Which host gets the most from it
These are judgement calls rather than findings, so read them as starting positions to test.
- A host with one or two properties and a calm calendar. The saving is small. Try the free parts of whatever you have and keep your own voice for the rest.
- A host with five to ten units and changeovers most weekends. The chore tracking (payments, cleaner timings, check-in instructions) is probably worth more than the message drafts. Test that side hardest.
- A host who already uses a separate guest-messaging tool. Find out whether the co-host duplicates it before paying twice. The comparison of guest-messaging tools for independent properties shows what is on the market.
- A host whose guests mostly write in another language. Test two threads in that language before anything else.
Checking after a month that it is earning its place
At the end of the four weeks, answer five questions. If you cannot answer one, you do not have the evidence yet.
- How many minutes a week did messaging take before and after?
- Did any guest comment on the tone of a message, good or bad?
- Did review scores or counts move? One month is too short to conclude much; note it and look again.
- How many D-scored drafts did you see, and were they all traced to a fixable note?
- What does it cost now, and what has Lodgify said it will cost after beta?
If the saving is real, the D count is falling and the price is known, extend it to more task types one at a time. If the price is still a mystery, keep the routines you can move: your property notes, your sample replies and your off-limits list all work with any tool. Choosing the right tool is only half the job; the other half is deciding which messages are safe to hand over, and that is something I help owners map in an AI implementation consultation if you would rather not work it out alone (contact).
Lodgify AI Co-Host: questions hosts ask before switching it on
Will Lodgify AI Co-Host send messages without my approval?
Lodgify says nothing is sent, changed or scheduled until you approve it, and that you can approve, adjust or skip each suggestion. It also says the co-host learns your preferences and can handle repeat tasks by itself later. Treat that later stage as something you grant deliberately, task by task, after the drafts have proved themselves.
How much does Lodgify AI Co-Host cost?
Lodgify's co-host page does not publish a price, and the feature is labelled beta. Check the pricing page and your account for which plan includes it, and ask Lodgify support whether the beta is free, bundled or priced separately, and whether that changes when beta ends. Get the answer in writing before you build routines on it.
Can it replace a human co-host or property manager?
Not on the evidence available. It drafts, reminds and tracks, but cleaners, repairs, safety problems and angry guests still need a person. Think of it as a fast assistant for the paperwork and routine messages. If your property manager mostly does messaging, test whether the co-host covers that part before changing the arrangement.
Is it only useful if I host through Lodgify?
The co-host works on your Lodgify account data, so it only knows what Lodgify holds: bookings, guests, messages, tasks. Lodgify's MCP connection lets compatible AI tools read the same context, which helps if you already use another assistant. If your bookings live elsewhere, a co-host with no data has little to work with.
Further reads
- Hospitable Airbnb Software: What Each Plan's AI Actually Does — See what a rival's AI replies do on each plan.
- Is HostBuddy AI Worth It for an Airbnb Host? — A different approach: a messaging-only AI add-on.
- Software for Vacation Rentals Pricing: How AI Sets Nightly Rates — Nightly-rate AI is a separate decision from messaging.
- Airbnb Cleaning Schedule: Coordinate Cleaners and Changeovers — The changeover chores a co-host is meant to track.
- How to Write B&B Room Descriptions and Guest Guides With AI — Write the guest guide the co-host will quote from.
- Holiday Let Website: Get Direct Bookings and Repeat Guests With AI — Direct bookings and repeat guests, the other half of the job.
- How to Automate Airbnb Messages: Test the AI Before You Switch It On — A pre-launch test script for Airbnb AI replies: wrong-answer traps, refund and safety requests, prompt tricks, pass marks and a staged switch-on.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Lodgify's AI Co-Host page and pricing page, read in October 2026. Plan names, beta status and pricing change; confirm them on Lodgify's own pages and in your account.