How to Build a 3-Day Itinerary by Chat With Real Prices

What do you need before building the itinerary?

You need three things: the client's message with their dates and preferences, Terra connected to your AI assistant via MCP, and a catalog of experiences loaded with real prices and availability. With that, a 3-day itinerary comes out of a chat conversation, no spreadsheet, no copying prices by hand.

Terra, Skyline's AI booking engine, builds day-by-day itineraries tailored to a traveler, with a public shareable link, and edits them keeping the same link. That's exactly what an operator needs when a client writes on WhatsApp and wants to see a proposal in minutes, not the next day.

Here's the full flow: from the first message to the final link, with prices calculated by the system, not by the AI.

How does the client's message become an itinerary proposal?

It all starts with what you already have: the WhatsApp conversation where the client says what they want to do, how many days they have and how many people are traveling.

Step 1: paste the chat or describe the request

If the client already wrote on WhatsApp, paste that chat into your AI assistant (Claude, ChatGPT or Gemini connected to Terra via MCP). Terra connects to these assistants like this: you generate a URL in Terra → Integrations, paste it into your assistant, and from there you operate by chat. You don't need to describe the request again in your own words if the client already wrote it out.

Example: an operator in San Andrés gets this message from a couple arriving Friday: "There are two of us, we arrive Friday afternoon and leave Sunday night, we like the beach but also some island history." That message, as is, is enough to get started.

Step 2: ask for recommendations before building anything

Before proposing a random day-by-day plan, check recommendations from the operator's real catalog. This avoids offering something that doesn't exist or has no availability that date. The assistant answers with options you can actually sell, not generic ideas from the internet.

Step 3: build the day-by-day plan

With recommendations in hand, ask the assistant to build a 3-day itinerary with those experiences, respecting arrival and departure times. Terra builds day-by-day itineraries tailored to the traveler, so the result isn't a loose list of activities but a plan ordered by date and time.

Where do the prices come from, and why does it matter?

Here's the point that worries an operator booking by chat the most: what happens if the AI gets a price wrong?

It doesn't happen, because it doesn't depend on the AI. Prices are computed server-side from the catalog; the AI cannot make up a price. This means that when the assistant builds the itinerary, every experience that appears carries the real value loaded in your catalog, with the rates, seasons or discounts you already defined there. If you change a price in the catalog, the next itinerary you build with that experience comes out with the updated price, not one saved in an old conversation.

For an operator, this has a practical consequence: you can delegate building the itinerary to the chat without reviewing every figure by hand before sending it. The verification already happened the moment the server calculated the number, not afterward.

What if the catalog has a price error?

If the error is in the catalog, the itinerary will reflect it too, because the server takes what's loaded there. That's why keeping the catalog updated is still your responsibility as an operator; what Terra solves is that the AI doesn't add a number on its own on top of that.

How do you share the link with the client?

Every itinerary you build has a public shareable link. That link is what you send to the client on WhatsApp, no need to generate a PDF or take a screenshot of anything.

The client opens the link and sees their 3-day plan with the experiences, times and price of each one. If the client wants to adjust something (change the day of the history tour, add an extra hour of snorkeling), you edit the itinerary and the link stays the same. This is key to not losing the thread of the conversation: you don't have to resend a new link every time something changes, and the client doesn't have to hunt through the chat for which was the "latest" valid link.

Step 4: edit without breaking the link

When the client replies with a change, go back to the chat with your assistant and ask it to update the existing itinerary: move an activity, change the time, add one more experience. Terra edits itineraries keeping the same link, so the client simply refreshes the page they already had open and sees the new version.

Example: the couple from San Andrés replies that they'd rather have the history tour Saturday morning instead of the afternoon. The operator adjusts that block from the chat, and the link already sent to the client shows the change without having to resend anything.

How does the itinerary become a payable booking?

An itinerary is the proposal. When the client says yes, the next step is turning it into something payable in one go.

Terra builds custom packages with a per-experience breakdown and a single payment link. That is, the same experiences that were in the day-by-day itinerary move into a package where the client sees how much each activity costs separately, but pays for everything with a single link, in one move.

Step 5: convert the itinerary into a package

From the chat, ask the assistant to convert the approved itinerary into a package. The system takes the day-by-day experiences and generates the per-experience breakdown along with the single payment link. That's the link you send the client to close the sale.

Step 6: send the payment link

The client receives the link, sees the breakdown of what they're paying for each experience in the itinerary, and pays from there. There's no intermediate step where you have to add up prices by hand or generate a separate charge for each activity.

Summary table: from message to payment

Why do this by chat instead of by hand?

The difference from building an itinerary in a spreadsheet or copying prices from memory is that here the price is never something you write, nor something the AI adds on its own: it comes from the catalog the moment the itinerary is generated. That reduces the risk of sending the client an outdated value, especially when you handle several experiences with seasonal rates.

If you also connect Terra to your assistant via MCP, the whole flow (recommend, build, edit, convert to package) happens in the same chat where you were already talking to the client, without switching tools halfway through the sale. If you want to understand in depth how that MCP connection works in Terra, AI Booking Engine for Tour Operators: Full Guide has the full detail on the protocol and the available tools.

What to do now

Generate your integration URL in Terra → Integrations, connect it to your AI assistant, and try the full flow with a real 3-day itinerary for the next client who writes to you on WhatsApp.