Upgrade to Pro — share decks privately, control downloads, hide ads and more …

The Science of Relevance: Turning Data Streams ...

Avatar for apidays apidays PRO
February 07, 2026

The Science of Relevance: Turning Data Streams into Customer Dreams

The Science of Relevance: Turning Data Streams into Customer Dreams
Speaker: Sebastien Grosjean, Founder & CEO at Smily

How do you turn raw data into human-centric travel experiences? In this session from TravelTech Paris 2025, Sebastien Grosjean (Founder & CEO of Smily) explores the shift from static personalization to contextual intelligence. Learn how to leverage real-time behavior and multi-source data signals to predict traveler preferences and deploy adaptive content at scale.

Key Themes:
• Contextual Intelligence: Moving beyond basic data points to "imagination at scale."
• Predictive Personalization: Identifying traveler personas before they book.
• Adaptive Content: Using GenAI to create dynamic, human-centric travel journeys.
• Conversion through Belonging: Strategies to transform data streams into emotional customer loyalty.

------------------------------------

Conference Details:
Conference: Travel Tech API Conference: Travel Unlocked! | part of FOST 2025
Date: 11 December 2025 • CNIT Forest, Paris

--------------------------

Resources from apidays:
Join our upcoming conferences: https://www.apidays.global/
Read the latest API news: https://www.apiscene.io

Avatar for apidays

apidays PRO

February 07, 2026
Tweet

More Decks by apidays

Other Decks in Technology

Transcript

  1. Family: built 100 Vacation Rentals + 2 Hotels past 39

    years Travel: 6+ months/year for 20 years Founded/Co-Founded 14 companies, mostly tech. Booking.com Connectivity Advisory Board Member Sebastien Grosjean
  2. Scalable All-In-One for Vacation Rentals 1.4B € in Gross Booking

    Value past 3 years World Best Partner of Airbnb, Booking.com and Expedia* * Nominated in 2021, still preferred partner or equivalent
  3. AI Adoption in Travel Source: Booking.com 50% 30% 18 to

    34 year olds have used AI to plan and research their vacations Over-65s have used AI to research a destination
  4. Trips Begin On Social Media use social media platforms (TikTok

    and Instagram) to make their travel decisions. 57% of Millennials and Gen Z travelers Source: Skift Research Exploring GenZ and Millennial Travel Habits 2023. Data as of November 2023
  5. Trips Begin On Social Media 75% of Millennials are influenced

    by social media when making their travel decisions
  6. 50x

  7. Promise The experience we commit to delivering. • Who are

    our ideal travelers? • Family ski holidays? Romantic getaways? Remote workers? • What experience do they get when they book with us instead of someone else?
  8. How we show up and how we express ourselves •

    Visual style: warm and cosy, minimalist and modern, playful and fun? • Tone of voice in your descriptions and guest communication • If your brand were a person, what would it look and sound like? Personality
  9. Why travelers and algorithms should believe us • Reviews: average

    rating and share of five-star feedback • User-generated content • Consistency everywhere: OTAs, website, social channels Proof
  10. Our brand is what AI and travelers infer from: From

    Listings To Brand • Photos • Descriptions • Reviews • Social content • Communication
  11. • Who are they? • Booking behavior • Key desires

    • Key pain points • Keyword bank • Landing pages • Social strategy • Exact experiences (quotes) • … AI can identify ideal travelers
  12. Experience Contextuelle de Smily Votre site personnalisera vos annonces en

    fonction de la saison et du client idéal détecté. À venir
  13. • Creating is easy • Accuracy is hard • Judge

    LLM / Content scoring agent / Evaluation agent is required 1 - Evaluation Agent is Required
  14. Cheaper models, run again and again, can outperform the fancy

    ones and still cost a fraction of the price. 2 - Cheaper Models
  15. 1. Source experts 2. Have experts create a playbook 3.

    Create the Evaluation Agent from the playbook 4. Create the Generation Agent 5. Feed real life data for reinforcement learning 3 - Agent Training Protocol