AI for Travel Agents Writes the Itinerary. The Quote Still Dies Unpaid

At ten past nine on a Tuesday night, Megan is at her kitchen table in Ballarat, finishing the fourth version of a three-week Japan itinerary for a family of five. Ryokan in Hakone, the rail pass maths, a day in Kanazawa because the eldest asked for it.

She sends it. The family thanks her warmly. Two weeks later she sees the daughter's photos from the same ryokan, booked direct.

Megan's software had a great year. An itinerary tool now drafts a day by day plan in minutes. A chatbot answers "do I need a visa" at midnight. A template turns her notes into a polished PDF with photos.

None of it noticed that the Japan quote was version four, that version one was where the client stopped being a buyer and started being a researcher, or that nothing was ever asked of them in return.

An agency gives away its expertise one quote at a time. No tool is counting what the quotes cost.

That is the gap in how AI for travel agents is being sold. Itinerary builders, content generators, chat widgets. All of it aimed at producing the quote faster. I build these systems for Australian businesses, so let me say the uncomfortable half out loud. A small agency rarely loses its year because quotes are slow. It loses it because quotes are free.

What AI for travel agents gets right

Give the itinerary tools their due.

Sequencing a route, checking that a train actually connects, drafting a hotel description that does not read like a brochure: that is bounded, repetitive work a model does well. Turning a two hour first draft into a twenty minute one is real.

But look at what it touches: the document. The industry builds there because a beautiful itinerary demos well, and nobody screenshots a quote conversion report at a conference.

Making the free part faster does not make it paid. It just lets you give away more of it per week.

The quote dies unpaid, not slow

Now walk the same month from the other end.

The honeymoon couple who wanted "just a rough idea" of the Maldives: three revisions, then silence.

The corporate client who asked for options on a team offsite, then booked through someone cheaper using your shortlist.

The cruise booked with a deposit in March, the final payment date in August that nobody flagged until the cabin was released.

Each one started as an email or a phone call. Each one is real hours or real commission. The remedy most independents already know about is a planning fee or a deposit before detailed work starts. It rarely gets applied, because asking for money feels harder than building one more version.

A hotel has the same disease from the supply side, holding rooms for a group that never confirms; I wrote about that half in AI for hotels.

The quote maths nobody runs

Run your own agency as arithmetic.

Say you build forty custom quotes a month at three hours each. That is a hundred and twenty hours. If fifteen of those book, the other twenty five took seventy five hours of skilled work and returned nothing.

Over a year that is nine hundred hours. Call your time sixty dollars an hour and the year has quietly absorbed fifty four thousand dollars of planning that someone else turned into a booking, or nobody did.

Swap in your own numbers; the shape survives any honest figure. Some unbooked quotes are simply the cost of being in business, and you should wear them. But moving even a third of that work behind a fee, or killing it at the first conversation, gives back weeks of your year.

The leak is not that clients are disloyal. It is that the agency never asked for a commitment before the expertise left the building.

Start with the enquiry, not the itinerary

Pick one workflow, not the whole agency.

A system reads every new enquiry as it lands, from the web form, the inbox and the phone notes. It asks the qualifying questions you would ask on a good day: dates, budget, who is travelling, whether they have booked with an agent before. It scores the enquiry and drafts your reply. A simple domestic trip gets a quick answer. A complex custom trip gets your planning fee terms and a payment link before version one is written.

You approve each reply on your phone with a yes or a no. Nothing goes out in your name without you.

Then point the same loop at the bookings you already have. Every deposit, every supplier final payment date, every fare rule change from an airline email, logged and flagged a week before it bites, with a drafted reminder to the client ready to send.

That gives you two numbers each month: quotes built, and quotes paid for or booked. The distance between them is your unpaid research department.

Automate the paperwork, never the advice

Here is the line that keeps this safe.

No system should tell a traveller whether they need a visa, what travel insurance covers, or whether a destination is safe to visit this month. Those answers change, they carry real consequences, and a model that answers confidently and wrongly will cost you a client and possibly a complaint. Point people to the official source and give your own judgement. The system qualifies, logs, reminds and drafts. The agent advises and decides.

Aim the software at the quote pipeline, never at the travel advice.

What AI for travel agents costs

Work down this list in order and stop the moment something works.

• The tools already inside your booking or CRM platform. Enquiry forms with required fields, automated deposit reminders, task flags on final payment dates. Often half configured. Turn them on first.
• An itinerary or content subscription, priced per seat each month. Cheap enough to trial on your next ten quotes and judge on hours saved.
• A custom build that reads your inbox and supplier emails, qualifies enquiries, drafts the fee terms and payment reminders, and writes back into the software you already run. A few thousand up to the low teens of thousands, depending on how many systems it must talk to.

Hold that against the unbooked quote hours above. What catches agencies out is the running cost rather than the build, and I broke those layers apart in what AI actually costs a small business.

The Australian layer: consumer law and passport data

Two things separate a system built for an Australian agency from an overseas template.

The first is consumer law. Your fee terms, your cancellation terms and what you say about a supplier's product all sit under Australian Consumer Law. A tool that writes your planning fee terms clearly, before work starts, protects you. One that invents a refund promise in a chat window creates the dispute you were trying to avoid.

The second is the data. A client file holds passport numbers, dates of birth, health notes for travel insurance and card details. That is the kind of information the Privacy Act takes seriously. Ask where the data is stored, whether it trains someone else's model, and whether you can get it out when you change platforms.

Common questions about AI for travel agents

What should a small agency automate first?

Enquiry qualification. It runs on enquiries you already receive, and you can measure it inside a quarter: quotes built against quotes paid for or booked.

Will AI replace travel agents?

Not the good ones. Travellers can already generate an itinerary for free. What they pay for is judgement, supplier relationships and someone to call when the connection is missed.

Should I charge a planning fee?

That is your call, and plenty of independents now do for complex trips. A system just makes sure the question gets asked every time, instead of only on the days you feel brave.

Go back to Megan. She did not have a bad month. She built four versions of a beautiful Japan trip and was paid for none of them. Her sent folder was keeping score all year. Nobody read it as money.

If you want a straight read on what your last quarter of quotes actually gave away, that is what a first call is for. Book a strategy call and bring your enquiry list and your bookings for the same three months. We will count the unpaid hours before we talk about building anything.