Units
1. Introduction to Artificial Intelligence in Tourism and Hospitality: Roles, Boundaries, Authentication and Privacy 2. Reservation and Channel Management: Occupancy, No-show and Overbooking 3. Revenue Management and Dynamic Pricing 4. Multilingual Guest Service: Chatbots, Translation and Response Drafts 5. Personalized Recommendations and Guest Experience: Upsell and Cross-sell 6. Tour, Content and Marketing Production: Statements, Social Media, Brand Voice 7. Guest Reviews and Reputation Management: Sentiment Analysis and Response 8. Front Office, Housekeeping and Operations: Demand Forecasting and Scheduling 9. Artificial Intelligence in Food & Beverage, Event and Facility Operations 10. Data Privacy, Cultural Sensitivity, Copyright and Brand Security 11. End-to-End Integration: Hotel and Agency Case Study and Roadmap
Unit 5 / 11

Personalized Recommendations and Guest Experience: Upsell and Cross-sell

Gains:

  • Ability to draft personalized room upgrade, additional service and experience suggestions based on guest segments with artificial intelligence
  • Ability to construct upsell and cross-sell offers in a language that does not disturb and respects personal data
  • Recommendations are subject to guest approval, actual stock and price; Understanding that artificial intelligence should not produce fake offers

Two guests come to the same hotel. One is a honeymoon couple; the other is a manager coming for a business meeting. Sending the same standard email, the same breakfast offer, the same spa brochure to both of them is a waste of opportunity. Offering a romantic dinner and late check-out to the honeymoon couple, and a quiet room and fast breakfast to the business guest will both satisfy the guest and increase revenue. That's personalization. In this unit, you will learn how to use artificial intelligence (AI) to create personalized recommendations, upsell (upsell — upgrade to a better/expensive product), and cross-sell (add a complementary product) offers based on guest segments. Critical principle: recommendations are based on actual stock, actual price and guest approval; AI does not create fake offers or disturb.

Difference between upsell and cross-sell

Let's clarify two concepts. Upsell's goal is to sell an upgraded version of what the guest gets: an upgrade from a standard room to a seaview suite, an upgrade from an economy tour to a private guided tour. Cross-sell is adding a complementary product to the existing purchase: adding a spa package to the room, dinner, airport transfer, tour ticket. The common goal of the two is to increase revenue per guest (ancillary revenue - additional non-room revenue); But it is essential to do this without disturbing the guest and adding real value to him.

Genre

Description

Hotel management example

Tour/agency example

upsell

Upgrade to higher version

Upgrade to suite

Private guided tour

cross-sell

Adding a complement

Spa + dinner

Museum ticket + transfer

Packaging

present together

honeymoon package

City + nearby destination

Segmentation: what is recommended to whom

The basis of personalization is segmentation: grouping guests based on common characteristics. Typical segments: honeymoon/couple, family, business traveler, group, extended stay, loyalty member. Each segment has different needs and willingness to pay. Once you give the segment and your actual list of services, AI can draft recommendation texts and offer combinations suitable for each segment. What it cannot do is to determine which guest is in which segment by diving into personal data and recommend a service that does not actually exist. Segment information comes from your CRM; service and price from your system.

Attention: Uploading personal data to AI to "analyze our guest list and tell whom to sell what" is both a privacy violation and risky. Instead, provide an anonymous segment description (“honeymoon segment”) and request recommendation text appropriate for that segment.

The art of proposal that doesn't bother

Poor personalization turns the guest into a sales target and overwhelms them. Good personalization delivers a relevant and easy-to-refuse offer at the right time. Principles: (1) value comes first — say what the offer brings to the guest first; (2) few and to the point—one or two relevant suggestions, not three; (3) no pressure — “if you want” language, easy no; (4) timing — pre-check-in upsell, during-stay experience recommendation; (5) transparent price — no hidden fees. AI can produce text that follows these principles, but you set the tone and frequency.

Step by step: personalized offer with AI

  1. Define the segment. Write an anonymous profile like "Honeymoon couple, 3 nights, no sea view."
  2. Give actual inventory. List only the services and prices you can actually offer.
  3. Request a quote. Draft 1-2 upsell/cross-sell suggestions suitable for the segment in a polite text.
  4. Verify. Confirm that each recommended service is actually available and priced correctly.
  5. Add the personal touch. Send it in a clear, unprinted form for guest approval.

three mini cases

Case 1 — Accurate upsell. A city hotel was sending an email two days before check-in offering a quiet floor + express breakfast upgrade to business segment guests. They personalized the text segment-by-segment with AI; They filled in the actual price and availability themselves. Upsell acceptance rate increased from 9% to 16% because the offer matched the guest's real need. AI wrote the text, human verified the inventory.

Case 2 — Fake service trap. An employee asked the AI ​​to "write down the activities we can recommend to the guest" and did not provide an inventory. YZ suggested a "rooftop infinity pool" and a "free boat tour" which the hotel didn't have. These were put on offer, the guest asked, there was no service and the hotel was embarrassed. Lesson: Asking the AI ​​for suggestions only without providing the actual list of services is an invitation to fabrication.

Case 3 — Disturbing frequency. One property was sending 3-4 sales messages to the guest every day throughout the stay; Guests were overwhelmed and satisfaction dropped. AI was asked to draft a single, well-timed experience proposal for the entire stay, following the principle of “short and to the point.” The number of messages decreased, acceptance rate and satisfaction increased together. AI simplified the text, the strategy came from the human.

Weak prompt / Strong prompt

Weak prompt:

Recommend anything we can sell to our guests in the way that will make the most money.

This prompt contains two errors: no actual inventory is given (the AI ​​makes it up), and the "most money" goal produces aggressive, disruptive bids.

Powerful prompt:

Your role: guest experience assistant. Purpose: low-pressure recommendation that adds value to the guest. Segment (anonymous): honeymoon couple, 3 nights, standard room. My REAL inventory (recommend only these): room upgrade suite +1,200 TL/night, romantic dinner 2 people 1,500 TL, late check-out at 14:00 free, spa couple massage 900 TL.Task: choose MAXIMUM 2 recommendations that fit this segment, give a warm and brief Write a draft email, use "if you want" language, make it easy to reject. DO NOT make up any services or prices that are not on my list.

Template: quote set by segment:

Segments: [family / business / couple / extended stay].My actual service list: [service: price, condition].Task: match 1 upsell + 1 cross-sell that best suits each segment and write a one-sentence justification. Do not suggest anything that is not on the list.

Template: upsell message before check-in:

Context: guest [segment], check-in [date], available room [type].Actual upgrade options: [option: difference price, availability].Task: draft a polite, no-pressure upgrade email.I will confirm availability and price; put [HERE] instead of blank.

Template: experience/activity suggestion:

Activities that I can actually offer: [list].Segment: [family, children ages 6 and 9].Task: suggest 2 activities suitable for this family, write in one sentence why they are suitable for each. Adding activities that are not on my list.

Common mistakes

  • Asking for suggestions without providing inventory. AI fabricates non-AI service; always give the actual list.
  • Uploading personal data and asking "who should we sell it to?" Breach of confidentiality; Use anonymous segment definition.
  • Aggressive "most money" goal. It disturbs, reduces satisfaction; Be value-oriented.
  • Offer very often. Less and more accurate always sells more.
  • Failure to verify price/availability. An incorrectly priced offer undermines trust.
Tip: Add the phrase “don't make up any services or prices that aren't on my list” in every quote prompt and double-check each quote against actual inventory before submitting. A fabricated offer will cause a guest to lose trust.

In summary

Personalization is about delivering a relevant, unpressured offer to the right guest at the right time; It increases both satisfaction and non-room income. AI accurately drafts upsell and cross-sell texts when you give the segment and actual inventory. But segment information comes from your CRM, service and pricing system; It should not recommend a non-AI service and should not overwhelm the guest. Recommendations are based on actual stock, actual price and guest approval.

Application task

Select a segment (e.g. "family, two children"). Write an inventory of 4 real or hypothetical services. With the “power prompt” template, ask the AI ​​for a no-pressure email draft with up to two suggestions suitable for this segment. Compare the output to your inventory to check for fake services and explain in 4 sentences why the offering fits this segment.

checklist

  • [ ] Have I only given the actual service and price list to AI?
  • [ ] I defined the segment anonymously, did I not upload any personal data?
  • [ ] Have I kept the number of suggestions small and accurate?
  • [ ] Have I verified the price and availability of each offer?
  • [ ] Can the language be rejected without pressure and easily?
Previous Next