Skip to content
NagentNagent
Log inSign upHire your AI Growth Team
NagentNagent
Log inSign up

travel · 8 use cases

Top agentic AI use cases for travel and hospitality in H2 2026

Travellers now plan trips with AI assistants and, in the US, can book a hotel without leaving Google, while new pricing, messaging, privacy and package rules arrive on fixed dates. These eight use cases respond to what changed in 2025 and 2026.

Which agentic AI use cases should travel and hospitality companies prioritise in H2 2026?Travel and hospitality teams should prioritise agentic work that follows the 2025 and 2026 changes: keeping rates and content ready for hotel booking inside Google AI Mode, being named in AI trip planning, one all-in price on every channel, direct offers in the EEA now parity has gone, AI disclosure on EU chat, WhatsApp under Meta’s new terms and DPDP consent records.

Rates, content and points ready for booking inside Google AI Mode

On 27 August 2026 Google began rolling out hotel booking inside AI Mode in the US, in English, with ten launch partners including Booking.com, Expedia, Marriott and Hilton; the hotel or booking platform stays merchant of record. NIA checks what AI Mode shows for your properties and keeps the rates, policies and room content it reads accurate.

Each week, NIA asks Google AI Mode the trip questions the group’s guests ask and records which properties appear and which booking options are offered. It finds a resort whose cancellation policy in Hotel Center no longer matches the booking engine. MIRA drafts the feed and content fixes. The distribution manager approves every change to rates, policies and room descriptions before it reaches the feed, and bookings from AI Mode become a tracked channel.
a scene from the weekDistribution manager at a US hotel group. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerEach week, run the trip questions guests ask through Google AI Mode for each property’s market.
  2. Record which properties appear, which booking options are offered and what price and policy are shown.
  3. Compare rates, taxes, fees, room descriptions and cancellation terms in Google Hotel Center with the booking engine.
  4. Draft fixes to the feed and room content, and check that member and points prices display correctly where the brand runs a loyalty programme.
  5. Approved by Distribution managerThe distribution manager approves every change to rates, policies and content before it is pushed.
  6. Push the approved changes and re-run the questions to confirm what AI Mode now shows.
in the roomAIMIRA with NIAPeopleDistribution managerTry this workflow
connects toGoogle AI Mode, Google Hotel Center, Booking engine and channel manager, Loyalty platform
measure it byProperties shown for priority trip questions, Bookings attributed to AI Mode, Feed and policy mismatches found and fixed

Guardrail. Google says the hotel or booking platform remains merchant of record and handles customer service, so every price, cancellation term and promise in the feed is yours to honour.

Recommended in ChatGPT, Gemini and AI Mode trip plans

Booking.com’s 2026 Travel Trends survey of 32,800 travellers found 69% of Gen Z and 64% of millennials use generative AI for travel planning. DRIS runs the trip-planning questions your guests ask across ChatGPT, Gemini, Perplexity and Google AI Mode, records which properties and sources are named, and drafts fixes to the pages and profiles the engines draw on.

Every week, DRIS asks ChatGPT, Gemini, Perplexity and Google AI Mode where to stay for a family beach holiday and a quiet anniversary trip in the resort’s region. A rival keeps being named for family stays, citing its kids’ club page. DRIS drafts a clearer family page, an FAQ on pools and babysitting, and an updated Google Business Profile. The SEO lead approves each change and its facts, and share of answers is now tracked.
a scene from the weekSEO lead at a resort collection. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerEach week, run the agreed trip-planning questions by source market and traveller type across ChatGPT, Gemini, Perplexity and Google AI Mode.
  2. Record which properties or operators are named, in what order, and which pages, profiles and review sites are cited.
  3. Compare with last week and with named competitors to find the questions you are losing.
  4. Draft fixes: plain answers on room types, accessibility, parking, family facilities and cancellation, plus profile updates on Google and the OTAs.
  5. Approved by Content or SEO leadThe content or SEO lead approves each change and its facts before publishing.
  6. Publish, then re-run the affected questions and log whether mentions and citations moved.
in the roomAIDRISPeopleContent or SEO leadTry this workflow
connects toChatGPT, Gemini, Perplexity and Google AI Mode, Google Search Console, Google Business Profile, WordPress or Webflow
measure it byShare of answers naming the property, Citations to your own site, Bookings from AI referrals

Guardrail. Fix the facts on the page rather than writing for a machine, and never publish a claim about rooms, accessibility, prices or awards the property cannot back up.

One all-in price across site, metasearch, ads and messages

Price rules tightened in 2025: the FTC’s fees rule took effect in the US on 12 May 2025 and covers short-term lodging, and the UK’s DMCC Act regime came into force in April 2025, banning drip pricing with fines of up to 10% of global turnover. India’s amended E-Commerce Rules follow from 1 January 2027. The agent audits every place a price appears and drafts fixes for approval.

Before the summer sale, MIRA crawls the booking engine, metasearch feeds, Google Ads copy and pre-arrival emails, and lists every price shown without mandatory fees. It finds a resort fee added only at checkout on one site and a sale banner comparing against a price that was never charged. LIPI drafts corrected templates. The director of e-commerce approves each fix, and the sale launches with one total everywhere.
a scene from the weekDirector of e-commerce at a hotel group in India and the UK. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerBefore each campaign, crawl the booking engine, landing pages, metasearch feeds, ad copy and email and WhatsApp templates for every displayed price.
  2. Flag prices shown without mandatory fees, fees added late in checkout, and paid extras selected by default.
  3. Check every sale or was-and-now claim against the price actually charged before the offer.
  4. Draft corrected copy, feed values and templates for each market’s rule.
  5. Approved by Director of e-commerceThe director of e-commerce approves each fix and the campaign’s price claims before launch.
  6. Apply the fixes, re-crawl and keep the audit record as evidence.
in the roomAIMIRA with NIA and LIPIPeopleDirector of e-commerceTry this workflow
connects toBooking engine, Google Hotel Center and Google Ads, Email and WhatsApp templates, Screaming Frog or another site crawler
measure it byPrice displays without mandatory fees found and fixed, Time from finding to fix, Checkout completion rate

Guardrail. From 1 January 2027 India’s amended E-Commerce Rules require e-commerce entities to run a yearly dark patterns self-audit and treat the prior price as the lowest of the previous thirty days.

free guide and deckEvery travel use case on these pages, with its workflow drawn, to share with your team.Get it free ↓

Direct offers in the EEA now that Booking.com parity is gone

Under the EU Digital Markets Act, hotels on Booking.com in the EEA can set their prices and conditions freely on other channels, including their own site, and a Commission factsheet of 28 September 2026 says outside prices no longer decide Booking Sponsored Benefit eligibility. MIRA plans member rates and direct offers by market, and the revenue manager approves where the direct price sits.

Spring rate planning opens. MIRA compares the direct rate with Booking.com and Expedia by market and length of stay, and proposes a member rate and a direct-only breakfast package for EEA guests. NIA drafts metasearch and search ads that carry the member rate. The revenue manager approves the rate positions and offers before they load, and direct share by market becomes the number the team reviews each month.
a scene from the weekRevenue manager at a hotel group in southern Europe. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerEach month, pull rates and availability by channel from the channel manager and the rate shopper.
  2. Compare direct, OTA and metasearch prices by market, date and length of stay for EEA properties.
  3. Propose member rates, direct-only packages and closed-group offers, each with its expected cost of acquisition.
  4. Draft the metasearch, search and email activity that carries each offer.
  5. Approved by Revenue managerThe revenue manager approves rate positions and offers before they are loaded.
  6. Load the approved rates, launch the activity and report direct share by market.
in the roomAIMIRA with NIAPeopleRevenue managerTry this workflow
connects toSiteMinder or another channel manager, Lighthouse or another rate shopper, IDeaS or Duetto, Google Hotel Center
measure it byDirect share of room nights by market, Cost of acquisition by channel, Member sign-ups

Guardrail. The DMA freedom applies to Booking.com in the EEA, so check every other OTA contract and market before pricing below an OTA, and show each member rate as an all-in total.

A website booking assistant that says it is AI, for EU guests

From 2 August 2026, Article 50 of the EU AI Act requires AI systems that talk to people to disclose that they are AI, unless that is already obvious. Convexa answers questions on rooms, policies and the area from approved content, says it is AI in its first message and hands over to reservations whenever a guest asks.

Late at night, a guest on the hotel’s site asks whether dogs are allowed and if the spa opens on Sundays. Convexa replies in German, says in its first line that it is an AI assistant, answers from the approved house rules and links the room with the right rate. When the guest asks about a wedding, it passes the chat to the events team. The reservations manager approves the answer library and every change to it.
a scene from the weekReservations manager at a lakeside hotel in the EU. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerA visitor opens chat on the website or booking engine.
  2. Open with a plain statement that the assistant is AI and that a person is available.
  3. Answer from the approved library of room facts, house rules, policies and local information, linking to the room and rate that fit.
  4. Hand over to reservations or events, with the conversation attached, when the guest asks for a person or raises a group, wedding or complaint.
  5. Approved by Reservations managerThe reservations manager approves the answer library and every change to it before it goes live.
  6. Review unanswered questions weekly and add approved answers.
in the roomAIConvexaPeopleReservations managerTry this workflow
connects toWebsite and booking engine, Approved answer library, Reservations inbox or CRM
measure it byChats that lead to a booking, Hand-over rate to reservations, Unanswered questions per week

Guardrail. Disclose the AI in the first message for EU guests under Article 50, and never let the assistant invent a price, policy or availability that is not in the approved library.

WhatsApp booking conversations under Meta’s 2025 and 2026 rules

WhatsApp changed twice for travel sellers: per-message pricing from 1 July 2025, with utility messages sent in an open service window charged from 1 October 2026, and terms from 15 January 2026 that bar general-purpose AI assistants while still allowing businesses that use AI to serve their own customers. LIPI runs enquiry and pre-arrival conversations inside those rules, with the CRM lead approving templates.

A family asks on WhatsApp about a villa for the Diwali long weekend. LIPI replies within the open conversation from the villa’s approved details, with the total including GST, and asks before adding the family to offers. SERA checks each outgoing template’s category, since Meta now prices messages one by one. The CRM lead approves every template and the opt-in wording, and bookings, opt-ins and opt-outs land in the CRM.
a scene from the weekCRM lead at a villa and homestay brand in India. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerA guest messages on WhatsApp from a click-to-WhatsApp ad, the website or a booking confirmation.
  2. Reply within the open customer service window from approved property details, with prices shown as totals.
  3. Classify each outgoing template as marketing, utility or authentication, and send marketing only to guests who opted in.
  4. Offer a person by chat transfer, phone or email whenever the guest asks or the question falls outside the approved answers.
  5. Approved by CRM leadThe CRM lead approves every template, its category and the opt-in wording before use.
  6. Log bookings, opt-ins and opt-outs to the CRM and report messaging cost per booking.
in the roomAISERA with LIPIPeopleCRM leadTry this workflow
connects toWhatsApp Business Platform, Gupshup, Twilio or another WhatsApp provider, Zoho CRM or HubSpot, Booking engine
measure it byBookings from WhatsApp conversations, Messaging cost per booking, Opt-out rate

Guardrail. WhatsApp’s policy requires opt-in before messaging, honoured opt-outs and a clear path to a person, and Meta’s 2026 terms allow AI that serves your own guests but not a general-purpose assistant.

Consent records for Indian guest data before the DPDP deadlines

India notified the DPDP Rules on 13 November 2025: consent manager registration applies from 13 November 2026 and the core duties, including standalone consent notices, from 13 May 2027. CORA builds a consent record for every guest, enquiry and lead across the booking engine, front desk, WhatsApp and CRM, and the data protection lead approves the notice text.

With the consent deadline approaching, CORA maps every place the chain collects guest data: the booking engine, the front desk, WhatsApp enquiries, wedding forms and the loyalty app. It finds forms that bundle marketing consent with the booking terms. SERA drafts standalone notices that state each purpose and how to withdraw. The data protection lead approves the notice text and consent fields, and every marketing list now carries a consent record.
a scene from the weekMarketing lead at a hotel chain in India. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerMap every point that collects guest or lead data: booking engine, front desk, WhatsApp, event enquiry forms, loyalty app and OTA imports.
  2. Flag forms that bundle marketing consent with booking terms or collect more than the stated purpose needs.
  3. Draft standalone notices that list the data, each purpose and how to withdraw consent or complain.
  4. Add consent fields with timestamp and purpose to the CRM, and suppress contacts without a valid record.
  5. Approved by Data protection leadThe data protection lead approves the notice text and consent fields before they go live.
  6. Report weekly on contacts with valid consent by purpose, and process withdrawals.
in the roomAISERA with CORAPeopleData protection leadTry this workflow
connects toPMS and booking engine, Salesforce or Zoho CRM, WhatsApp Business Platform, Loyalty app
measure it byContacts with a valid consent record by purpose, Time to process a withdrawal, Collection points with a standalone notice

Guardrail. Rule 3 of the DPDP Rules 2025 requires clear, standalone notices that name the data and its purposes and give a simple way to withdraw consent, from 13 May 2027.

Trip bundles and add-ons checked against the revised EU package rules

The EU’s revised Package Travel Directive, (EU) 2026/1024, entered into force on 28 May 2026: it removes linked travel arrangements, simplifies the package definition and gives member states twenty-eight months to transpose, with full application expected from 2029. FOLIO reviews every bundle and cross-sell flow and drafts the changes for the compliance lead to approve.

Planning for the new package rules starts with an inventory. FOLIO lists every flow where a traveller can add a hotel, transfer or tour to a flight, and every bundle sold at one price. CREA drafts the information each flow will need and flags cross-sell emails that could turn a sale into a package. The compliance lead approves the changes and the timeline, and the product team builds from one list.
a scene from the weekProduct manager at an online tour operator in the EU. Illustrative, built from this workflow, not from a client.

the workflow, in one room

  1. TriggerList every bundle, dynamic package and cross-sell flow on the site, in apps and in post-booking emails.
  2. Mark which flows sell more than one travel service for the same trip, and how the traveller is invited to add each one.
  3. Compare each flow with the revised package definition and the new complaint-handling duties.
  4. Draft changes to information pages, booking flows and emails, with owners and dates set against each national transposition timeline.
  5. Approved by Compliance leadThe compliance lead approves the changes and the timeline before build starts.
  6. Track each change to release and keep the record for regulators.
in the roomAICREA with FOLIOPeopleCompliance leadTry this workflow
connects toBooking platform and dynamic packaging engine, Email platform, Jira or Asana, Insolvency protection provider
measure it byFlows reviewed against the new rules, Changes shipped before national deadlines, Complaints acknowledged on time

Guardrail. The revised directive requires organisers to acknowledge complaints within seven days and answer within sixty, and national laws set the final detail, so confirm each market with counsel.

Frequently asked questions

Can AI agents book hotels for travellers yet?

In the US, yes: since 27 August 2026 Google has been rolling out hotel booking inside AI Mode with ten partners, including Booking.com, Expedia, Marriott and Hilton, with payment by Google Pay. The hotel or platform stays merchant of record. For everyone else, the work is making rates, policies and content accurate wherever AI assistants read them.

What changed for hotel and travel pricing rules in 2025 and 2026?

The FTC’s fees rule took effect on 12 May 2025 and requires US short-term lodging to show the total price with mandatory fees. The UK’s DMCC Act regime banned drip pricing from April 2025. In India, amended E-Commerce Rules from 1 January 2027 add a yearly dark patterns self-audit and a thirty-day prior price rule.

Do we have to tell guests they are chatting with an AI?

For people in the EU, yes, from 2 August 2026 under Article 50 of the EU AI Act, unless it is obvious from the context. Saying so in the first message is the simplest way to comply in every market, and it lets guests ask for a person straight away.

Which of these should we start first?

Start with the work that has fixed dates: one all-in price on every channel, AI disclosure on any chat assistant, and consent records ahead of India’s May 2027 deadline. Track AI Mode and AI trip-planning answers alongside, because you cannot fix what you are not measuring.

sources

  1. Google: Book hotels with AI Mode in Search (27 August 2026)
  2. Statista: Gen Z consults gen AI for travel advice (Booking.com 2026 Travel Trends)
  3. FTC: Rule on Unfair or Deceptive Fees to take effect on May 12, 2025
  4. CMA: New consumer protection regime comes into force (2025)
  5. PIB: Government amends E-Commerce Rules, in force from 1 January 2027 (2026)
  6. European Commission: How the DMA ensures businesses using Booking.com are free to set their prices (2026)
  7. Cooley: EU AI Act transparency obligations take effect 2 August 2026 (2026)
  8. TechCrunch: WhatsApp changes its terms to bar general-purpose chatbots (2025)
  9. Meta for Developers: Pricing on the WhatsApp Business Platform
  10. WhatsApp: Business Messaging Policy
  11. SCC Online: MeitY notified Digital Personal Data Protection Rules, 2025 (2025)
  12. European Commission, Transition Pathways: Revised Package Travel Directive now in force (2026)

free guide

18 AI agent use cases for Travel & Hospitality

Every workflow on these pages as a designed guide and a presentation deck, ready to share with your team: a scene from the week, the workflow drawn with its approval gate, the systems it connects to and how to measure it. Light and dark editions, updated every month.

  • Top 10 sales and marketing use cases for AI agents in travel and hospitality
  • Top agentic AI use cases for travel and hospitality in H2 2026
  • Who is in the room: your AI team for travel and the people who approve
  • The guardrails and the rules they come from, with sources

We email you the link too, and may follow up once. Unsubscribe in one click.

Put these travel workflows to work. Start a team on your own data today, or book a session with the Agentic AI Lab and we map your first three workflows with you.