Superintelligence for Incoming Tourism Group Itineraries With Approvals

Superintelligence for Incoming Tourism Group Itineraries With Approvals

How to log and approve group itinerary changes in incoming tourism with AI before notifying guides and drivers, with clear limits and an audit trail.

•

Want to apply this in your company?

Book a call with Juan →

Superintelligence, in the sense of a system that surpasses human ability at practically everything, is not what runs a group itinerary today. What works in incoming tourism is an AI agent with clear rules. It receives the requested change (a swapped visit, a new time, a group that must arrive earlier), records what changed, checks what the rules allow, and only notifies the guide and driver after a responsible person approves. Without that record and that approval, a well-written message can take apart an entire group's day.

Key takeaways

  • A route change is an operational decision, not a conversation. It needs to show who asked, what changed, and who approved it.
  • Guides and drivers only receive instructions from an approved version of the itinerary, with the version date and time attached.
  • The agent can prepare the change proposal, but it does not confirm a booking, ticket, or commercial term on its own.
  • Passenger data stays at the minimum necessary, with restricted access and a retention rule defined together with legal counsel.
  • Every step is logged, so the operation can later explain what happened and why.

Why group changes need a record before the notice

In a group tour, a change is rarely just one line. Moving a visit from the afternoon to the morning affects arrival time, the venue's availability, transport, lunch, and sometimes what the client has already agreed to. When information travels through loose messages, each person receives a different version: the guide hears one thing, the driver another, and the client gets a third.

So the first job is not to automate the notice. It is to define where a change starts and where it is recorded. The request enters through a single channel, becomes a record with an identifier, and passes a check before anyone in the field is contacted.

Who approves what: a decision guide

Not every change needs the same person. A simple rule helps separate what the agent can prepare, what a coordinator approves, and what leaves the automated flow.

  • Adjustments to visit order within the same area: the agent proposes and the operations coordinator approves.
  • Swapping a venue or attraction with a booking already made: the agent lists options, and whoever manages the booking confirms with the supplier before any notice goes out.
  • A change that alters the price, cancels part of the service, or modifies commercial terms: it leaves this flow and goes to the sales team, without exception.
  • A time change that affects contracted transport: it requires confirmation from the transport supplier before the driver notice is released.
  • A passenger request without clear identification or outside the agreed schedule: it waits and is handled by a person.

Step by step: from request to notice for the guide and driver

This is the sequence we usually propose as a starting point for a pilot. It is simple to describe and demands discipline in execution.

  • Receive the request through a single channel, with the group and the current itinerary identified.
  • Open a record with the text of the request, who made it, the time, and the itinerary version before the change.
  • Check whether the change fits the rules: minimum lead time, area, type of change, and whether the requester has permission.
  • Build the proposed new version, highlighting what changes and what is affected: transport, times, meals, and suppliers.
  • Send the proposal to the approver, with an explicit answer of approve, adjust, or decline.
  • After approval, record the approved version, along with who approved it and when.
  • Only then send the notice to the guide and driver, with the approved version and only the data each one needs.
  • Record the delivery confirmation. If it does not arrive within the agreed time, the operation receives an alert to call.

Illustrative scenario: a visit swap on a rainy day

Illustrative scenario: a group has an outdoor park visit scheduled for the afternoon. Based on the forecast, the field team considers moving the visit to the morning and shifting lunch. Someone in coordination logs the request through the operations channel.

The agent opens the record, notes the current itinerary version, and finds that the change affects lunch, which is booked with the restaurant. It does not confirm anything with the restaurant. It builds the proposal with the new timing, marks the point that depends on the supplier, and sends it to the coordinator. The coordinator approves the earlier visit and asks that the restaurant contact be reached before any notice goes out.

Only after the restaurant confirms does the agent send the guide and driver the new version, with departure time and meeting point. Each one receives only what they need to know. The record keeps the original request, the approval, the supplier confirmation, and the time of the notice. If the driver does not respond, the operation receives an alert to call. The gain is less about speed and more about preventing someone from leaving with a version the operation had already abandoned.

What the change record needs to contain

A useful record answers the questions that come up later: who asked, what changed, what was approved, who confirmed, and when. It does not need to be complicated.

  • Group and itinerary identifiers, with the current version.
  • The original text of the request and the channel it came through.
  • The changed fields, shown as before and after.
  • The decision of the approver, with their role and the time.
  • The supplier's response when the change depends on them.
  • Messages sent to the guide and driver, with send time and read time when the channel provides it.
  • The reason for any decline or adjustment.

The idea of tracking every step of a flow, described in observability concepts, applies here. Without a trail, there is no way to review a decision or adjust the rule afterward.

Passenger data and messages: LGPD and WhatsApp

Group itineraries carry personal data: names, documents when a venue or transport requires them, dietary restrictions, health conditions, and contact details. The agent should receive only what each task requires. The driver does not need the full list with every passenger's contact. The guide may need dietary restrictions, but rarely needs documents.

It is worth aligning these rules with whoever handles the legal side of the company and following the guidance of Brazil's data protection authority (ANPD) on personal data processing. On WhatsApp, the message to the guide and driver is operational, but the message to a passenger follows rules on conversation windows and approved templates. Before designing this flow, check what the 24-hour WhatsApp API window allows the agent to do.

How to start without losing control of the operation

The path we follow is to map, limit, pilot, and hand over. First, map how an itinerary change happens today, including the shortcuts the team already uses. Then limit: the agent reads and proposes, but does not confirm suppliers, change commercial terms, or send notices without approval.

Next, pilot one real type of change, such as visit order adjustments within the same area, with a measure agreed on before starting, for example the time between the request and the approved notice, and the share of notices that had to be corrected afterward. Finally, hand the process to the team and decide, using the pilot's own data, whether to expand, adjust, or stop. There is no promised number in this process. The operation chooses what to measure and records the result. If you want to see how this work is run with companies, read how we work with companies.

Frequently asked questions

What does superintelligence in incoming tourism mean in practice? Today the term works more as a horizon for discussion than as a tool you can use to run a group. Day to day, what exists are AI agents with defined tasks, approval rules, and decision records. To separate the concept from the product, it helps to read about the difference between ChatGPT and superintelligence.

Can an AI agent change an itinerary on its own? It is not advisable for it to confirm changes with suppliers, commercial terms, or notices to field staff without human approval. The right role for the agent is to receive the request, record it, check the rules, and prepare the proposal. The final decision should stay with a named person in the process.

How do you keep guides and drivers from receiving different versions of the same day? The main rule is that every notice comes from an approved version of the itinerary, with an identifier and a time. Loose messages should be replaced by sending the approved record, and any adjustment creates a new version. That way, the people in the field always know which version is valid.

What should the change record store, and who should have access? The record should store the original request, the approval, the supplier confirmations, and the notices sent. Access should be limited to the roles that need that information, and the retention period should be defined with legal counsel in line with the LGPD.

Sources