Why travel websites?

Low engagement.
Lack of relevance.
Lots of data.

You’re booking your next holiday or travel adventure, and the itinerary is coming together. You’re excited, right? Until you hit the checkout page. Suddenly, all that excitement? Gone. Generic, boring, lifeless. “your purchase is complete”. These words should add magic, not kill it.

These companies know a lot about you, where, when, why your traveling. But Personalisation? Practically nonexistent. At least, until now.

Problem Area

We noticed personalisation in travel stops the moment you hit ‘book now’—but that’s where it should start

Click to expand

Defining the Gap in Post-Purchase Personalisation

While most travel platforms pour their energy into personalising search and booking, the moment a trip is confirmed, the experience flattens. Across Expedia, Booking.com and even newer AI tools, personalisation stops at the transaction. Everything that follows — confirmation screens, checkout details, pre-trip nudges — becomes generic, despite companies already holding rich contextual data about the traveller. This disconnect became the foundation of the problem space I explored.

For this sprint, we weren’t aiming to reimagine the entire booking journey. Instead, I focused specifically on the post-purchase phase—moments like the checkout screen, confirmation page, and pre-trip reminders.

Brainstorm for AI-Driven Personalisation areas of possible integration

To ground the problem space, I mapped out all the touchpoints where AI could add contextual value using data platforms already collect — from marketing messages to after-purchase support. The diagram below captures that landscape. From this broader field, I deliberately narrowed the scope to a single, high-impact moment: the post-checkout trip confirmation page. It’s the first point where excitement peaks, personal context is richest, and current platforms still default to generic, boilerplate content.

Research

With no time for user interviews, we flipped the script and used existing data as our playground

Click to expand

Research Overview:
Understanding what the landscape is like (data, user journey, and expectations).

Because this sprint was about rapid AI exploration, there wasn’t time for extensive primary research. Instead, I leaned into desk research and focused on one question: what data do travel companies already have, and why isn’t it being used to personalise the moments after booking?

To map this out, I sketched the Travel Data Ecosystem — a visual breakdown of the user information already sitting inside most booking platforms: identity data, trip context, behavioural signals, partner integrations, preferences, and inferred insights.

Travel Data Ecosystem diagram

From mapping the Travel Data Ecosystem and reviewing how major platforms handle post-purchase interactions, a set of clear insights emerged:

Research Insights

There’s more than enough data already.
‍So the issue isn’t data scarcity, it’s dormant data. Travel companies already hold huge rich contextual information, but almost none of it gets activated after purchase. (It’s used extensively in recommendation engines pre-purchase ect)
Personalisation follows the transaction, not the experience arc.
‍Data is heavily used to optimise search, sort, and pricing, but barely used to shape the emotional arc of the journey (anticipation, reassurance, preparation) once the booking is confirmed.
Post-purchase touchpoints are treated as admin, not experience moments.
‍Confirmation screens, emails, and reminders are designed for verification, not value

Existing Solutuons

Everyone’s personalising the search—nobody’s personalising the journey after you’ve booked

Click to expand

Existing solutions:

How the industry currently tries to solve things (solution patterns, not competitors).

Solution Type

What It Does Well

Where It Falls Short

Smart Pricing Algorithms

Optimise prices in real time based on demand, user behaviour, and inventory.

Entirely focused on transactional optimisation, not on improving the traveller’s experience.

Package & Add-On Recommendations

Suggest bundles (tours, insurance, upgrades) based on past bookings or broad preferences.

Often generic and sales-driven, with little sensitivity to the specific context of the current trip or traveller.

AI Chatbots & Assistants

Handle support queries, FAQs, and simple changes through conversational interfaces.

Reactive by design: they only help if the user knows what to ask and chooses to engage.

AI-Generated Itineraries & Trip Builders

Auto-generate suggested activities, routes, and day plans.

Typically built from general content + search history, not from deeper lifestyle, confidence level, or accessibility needs.

Rule-Based Post-Purchase Emails & Alerts

Send confirmations, reminders, and generic pre-trip checklists.

One-size-fits-all messaging; little adaptation to who the traveller is or what situation they’re in.

Opportunity Areas

From these patterns, several gaps became clear:

Automation is optimised for revenue, not confidence.
Smart pricing, recommendations, and funnels are finely tuned to conversion and upsell metrics, but not to reducing uncertainty or increasing readiness for the trip.
AI is treated as a feature, not a layer.
When AI appears, it’s usually packaged as a visible tool (chatbots, “Ask AI”, magic buttons) rather than a quiet layer that reshapes content and behaviour in the background.
Post-purchase tooling is administrative, not experiential.
Existing systems handle tickets, changes, and alerts well, but do almost nothing to adapt content based on who the traveller is, why they’re going, or what they might need next.

Competitor breakdown

To position this concept, I analysed how major travel platforms currently use data and AI.

🗂️ What They Already Do Well

Booking.com

Airbnb

Google Travel

Expedia Group

Personalised search & filters

✅ Deep filtering, deal surfacing

✅ Strong intent-based filters

✅ Aggregated search + smart sorting

✅ Broad filtering across brands

Behaviour / history-based recommendations

✅ Recommends similar stays & destinations

✅ Learns from past trips and saved stays

✅ Uses search + browse history

✅ Uses loyalty + history to push deals

Dynamic pricing & conversion optimisation

✅ Aggressive price testing, urgency cues

✅ Smart pricing tools for hosts

✅ Highlights “best times” and price trends

✅ Heavy emphasis on discounts & bundles

Cross-platform / ecosystem data use

❌ Mostly internal

❌ Data siloed per listing

✅ Strong tie-in with Maps, Gmail, Flights

✅ Shared data across Expedia brands

Reactive support (chat, FAQs, help flows)

✅ Chatbot + help centre

✅ Messaging with hosts + support

✅ Surface help content contextually

✅ Chat + help flows across products

Where My Concept Diverges

Booking.com

Airbnb

Google Travel

Expedia Group

Personalised post-purchase experience

❌ Static confirmation pages

❌ Host-led info only

❌ Itinerary view, not tailored guidance

❌ Generic confirmation + loyalty copy

Proactive pre-trip support

❌ Mostly notifications for changes

❌ Depends on each host

~ Some auto prompts (check-in, routes)

❌ Limited to booking changes

Invisible AI (not chatbot)

❌ AI is surfaced as a help bot

❌ No embedded AI layer

~ Light AI summaries, still separate

❌ Interruptive prompts, not woven in

Emotional, humanised confirmation tone

❌ Purely functional

✅ Sometimes human via host messaging

❌ Instrumental tone

❌ Corporate, deal-first language

Personalisation beyond upsell

❌ Optimised for conversion

✅ Manual, story-driven but inconsistent

❌ Ad- and visibility-driven

❌ Focus on bundles and upgrades

This gap is exactly where my concept sits: using AI to personalise the moments after purchase; the confirmation, the lead-up, and the support no platform currently designs for.

Core Design Insights

Invisible AI > Visible AI. AI shouldn’t always be a chatbot or a button, it can quietly shape what the page says and shows, based on who’s looking at it.
Post-purchase is prime real estate. The confirmation and pre-trip moments are emotionally loaded and under-designed; they’re the best place to re-invest personalisation.
Activate existing data, don’t collect more. The opportunity isn’t new data pipelines, it’s using what platforms already know (age, trip purpose, destination, timing, group) to generate context-aware, personalised content at scale.

Bellow is a fully functional version of the prototype i’ve integrated into webflow. Try it yourself and see how it custom adapts to specific users and cases. The goal isn’t just dynamically generated website content, its hyper personalised content.

the product

Please wait for response to load.

This may take up to 2-3 mins.

Heading

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Your Tokyo family holiday is confirmed!

Hi Dewi — your Tokyo family holiday is all set!

It’s your first overseas trip, with seven nights in late December coming from Jakarta. Expect crisp winter days, dazzling lights, and kid-friendly adventures across theme parks, shopping malls, and relaxed sightseeing. Here’s a short, practical wrap-up tuned to your plans so you can travel with confidence.

Smooth arrival and entry
Most Indonesian passports require a Japanese tourist visa, unless you have an e-passport that’s been pre-registered for visa-free entry — confirm with the Japanese Embassy in Jakarta and allow time for processing. Before you fly, complete Visit Japan Web to generate immigration and customs QR codes; it’s optional but usually shortens the queue.

With kids and luggage, the Airport Limousine Bus is the easiest option from Haneda or Narita to major hotels and the Tokyo Disney area. Alternatively, choose the Keisei Skyliner (Narita) or Tokyo Monorail/Keikyu (Haneda) for a fast rail trip into the city. Tokyo is two hours ahead of Jakarta, so jet lag is minimal — plan early starts and earlier bedtimes on your first day.

Comprehensive travel insurance covering medical care and delays is strongly recommended for families.
Winter-ready packing and family days out
Late-December Tokyo is cool and dry, around 3–10°C, with sunset near 4:30 pm. Pack layers (thermal tops, jumpers, warm jacket) plus beanies, gloves, and comfortable walking shoes. Indoor spaces are well heated — dress so you can peel off layers easily, and bring lip balm and hand cream for the dry air.

For the parks, buy dated tickets in advance via official apps and check for same-day paid express options. Stroller hire and lockers are available at both Tokyo Disneyland and DisneySea.

Cosy, indoor picks for chilly days include Sanrio Puroland, the Sumida Aquarium at Tokyo Skytree, and easy shopping at Ikspiari (Disney Resort) or Odaiba’s DiverCity.
Getting around, money and easy eats
Pick up a Suica or PASMO transport card on arrival. Ask station staff to set up child cards for half fares — they can be used for trains, buses, and many shops. Japan is largely cashless, but smaller eateries still prefer cash — withdraw yen at 7-Eleven or Japan Post ATMs and use your card or phone where accepted. Stay connected with an eSIM or pocket Wi-Fi from the airport; having maps and translation tools handy makes your first trip simple.

For mild, family-friendly meals, try chicken or shoyu ramen, Japanese curry rice, tempura, and gyoza. If you need halal options, the Halal Gourmet Japan app is a reliable guide around Asakusa and Shinjuku.

Power sockets: Type A, voltage 100 V — bring a universal adaptor; most phone and laptop chargers (100–240 V) work fine.

Have a magical week in Tokyo — selamat jalan, Dewi!

Hidden Logic

Priority Topics:
Entry & arrival — first-time traveller with children from Indonesia; visa clarity is essential.
Included Visit Japan Web for efficiency, and family-friendly transfer options (bus, simple rail).
Weather & packing — key for a tropical family entering Japanese winter; included daylight timing and park prep.
Getting around, money, connectivity & food — matched to Dewi’s interests: easy sightseeing, malls, mild local food.
Added IC cards with child fares, cash/card mix, eSIM/pocket Wi-Fi, halal options, plug/voltage info.

Deprioritised:
JR Pass (not cost-effective for a Tokyo-only week), nightlife, onsen etiquette, temple manners, emergency numbers, vaccinations, coworking spaces, complex day trips (Hakone/Nikkō), and detailed Christmas/New Year closures (trip ends before New Year).

Assumption:
Dates noted as both January and 20–27 December; assumed late December (7 nights) for weather and crowd guidance.Pick up a Suica or PASMO transport card on arrival. Ask station staff to set up child cards for half fares — they can be used for trains, buses, and many shops.

Your Kyoto spring getaway is confirmed, Jordan!

Jordan, your Kyoto spring getaway is locked in.

Twelve nights gives you time to wander slow — perfect for temple walks, design detours, and third-wave coffee. April brings mild days, cool evenings, and the tail-end of cherry blossoms with fresh greens — ideal light for photography.

As a frequent traveller, this guide stays tight: key arrival moves, crowd-savvy timing, and design-forward picks to match your style.

Touchdown to town, the smooth way
Flying into Kansai (KIX)? Take the JR Haruka Express straight to Kyoto Station — about 75 minutes. Reserve a seat online and board with a QR code. If you arrive via Itami (ITM), the airport limousine bus is the most direct route to Kyoto Station.

Add an IC transit card (ICOCA, Suica, or PASMO) to your Apple/Google Wallet for tap-on fares across trains, subways, and most buses; top up in the app or at stations to skip ticket machines.

Within Kyoto, favour trains and walking — central buses are often slow in traffic.

Carry a small float of yen for shrines and smaller cafés (7-Eleven ATMs accept foreign cards), and consider completing Visit Japan Web to speed immigration and customs.
April rhythm: design, light and breathing space
Beat the crowds by going early:
• Fushimi Inari before sunrise,
• Arashiyama’s bamboo grove before 8 a.m.,
• Higashiyama lanes on weekday mornings.
Save blue hour for Yasaka Pagoda and Shirakawa-dori.

For architecture, mix contrasts — Hiroshi Hara’s Kyoto Station Skyway with wooden machiya streets, Tadao Ando’s Garden of Fine Art, and the KYOCERA Museum of Art with MoMAK for contemporary works.

If you have a spare day, the Miho Museum (by I. M. Pei) blends glass, steel, and hillside forest beautifully.
For coffee precision: Weekenders Coffee, Kurasu Kyoto, % Arabica (Higashiyama/Arashiyama), and Vermillion near Fushimi Inari.
‍
If your stay touches Golden Week (from 29 April), pre-book museums and restaurants.
Pack smart, move thoughtfully
Expect 8–20°C with light showers — pack layers, a compact waterproof, and shoes that handle cobbles and temple stairs. Many temples ban tripods and drones; in Gion’s private alleys, photography is restricted — ask before shooting people and keep gear discreet.

Solo dining is normal — book kaiseki counters and design-forward cafés ahead via TableCheck or a hotel concierge (strict no-show policies apply). Tipping isn’t customary; cards are widely accepted, though cash may still be required at smaller spots.

Kyoto in spring awaits — have an inspiring trip, Jordan!

Hidden Logic

Topic Selection:
- Skipped detailed visa guidance (Australians are visa-exempt for short tourist stays), SIM/eSIM brand comparisons, JR Pass cost breakdowns (12 nights centred on Kyoto rarely justify it), airport terminal maps, and safety/scam lists — low value for a frequent solo traveller.

Emphasis Choices:
- Prioritised fast arrivals and frictionless movement (Haruka/limousine, mobile IC cards, Visit Japan Web).
- Added crowd-smart timing and blue-hour/sunrise slots aligned to photography interests.
- Highlighted design/architecture and specialty coffee to match UX/design background and mid-high budget.
- Focused on April weather prep, cash nuance, and local etiquette (tripod/drone, Gion restrictions) to protect experience and show cultural respect.

Your Paris escape is confirmed, Harold & Margaret!

Bonjour Harold and Margaret — your 9-night Paris escape this September is locked in.

While it’s early spring in Melbourne, Paris enjoys mellow late-summer days, ideal for art, gardens, and unhurried strolls. As retired teachers returning to Europe after some time, this plan keeps things comfortable and curated without fuss. Below are the details that matter most — so you can settle in quickly and spend more time savouring than queuing.

Arriving with ease
After the long overnight from Melbourne, the easiest arrival at Paris Charles de Gaulle (CDG) is a pre-booked chauffeur: meet-and-greet at the exit, help with luggage, and a fixed fare to your hotel. Official taxis are also reliable with set fares — about €53 to the Right Bank or €58 to the Left Bank. Join the signed taxi queue and ignore touts. Expect 45–60 minutes to reach central Paris depending on traffic. The RER B train is an option, but with luggage and occasional stairs, a car is usually more comfortable.

To ease the 8-hour time difference, spend your first afternoon in gentle daylight — a riverside stroll and early dinner help the body clock adjust.
Art, gardens and viewpoints — without the queues
Book timed entries for major museums — especially the Louvre, Musée d’Orsay, and the Orangerie — and note weekly closures:
- Louvre — Tuesday
- Orsay — Monday
- Orangerie — Tuesday
The Paris Museum Pass can be good value if you plan several visits across your nine nights; even with it, reserve time slots for the busiest venues.

For beautiful, easy walks, visit the Jardin du Luxembourg and the Tuileries. Consider a half-day to Monet’s Garden at Giverny (open roughly April–November) — it offers level paths and gorgeous blooms. For accessible views, try the Montparnasse Tower or the Eiffel Tower lifts; in Montmartre, use the funicular to reach Sacré-Cœur comfortably.
Dining and daily practicalities, tailored to you
September is mild — around 12–22°C with the occasional shower. Pack layers, a light rain jacket, and supportive walking shoes.

Dining in Paris is reservation-led: book popular bistros or Michelin picks early, and look to lunch tasting menus for excellent value. Service is included, so tipping is optional — round up or leave 5–10% for standout service. Cards and contactless payments are accepted almost everywhere, but carry a little cash for markets.

Australia uses Type I plugs; France uses Type C/E (230V) — bring a universal adaptor. If you need data, set up a travel eSIM before you fly for reliable coverage.

Bon voyage — Paris is yours to savour, Harold and Margaret!

Hidden Logic

Priorities:
- Focused on arrival comfort (private transfer vs taxi, CDG flat fares, jet-lag management) — important for a long-haul couple with a preference for ease.
- Highlighted museum booking strategy and closure awareness due to strong art interest and likely September crowds.
- Suggested accessible garden walks and viewpoints suited to age and pace; added Giverny as a gentle, scenic day trip.
- Covered packing, reservations etiquette, tipping norms, payment, plug type, and data setup — all high-utility for Australians returning after a gap.

Omissions:
- Skipped detailed metro/Navigo advice (complex, less relevant given comfort preferences).
Left out scam/safety lists to keep tone calm and realistic.
- Omitted visa details since Australians are visa-exempt for short stays, but reminded to confirm rules before travel.
- Avoided specific festival/event mentions (like European Heritage Days) to prevent inaccuracies without firm travel dates.

Output / Outcome

So as you can see it generates something. But the biggest question is does it actually personalise the content to each specific user? is it providing valuable insights, advice or comfort to the users? or is it generating generic content each time?

To evaluate this, I created a persona checklist — a structured way to test whether the AI was genuinely adapting its tone, content, and priorities to each user, rather than just generating generic travel copy. Each checklist outlined what I expected to see (key inclusions and omissions) based on that traveller’s background, experience, and goals. By comparing the generated output against these criteria, I could measure how well the model understood contextual nuances — for example, whether it gave family-focused reassurance for Dewi, efficient insider tips for Jordan, or comfort and accessibility for Harold and Margaret. This process turned subjective impressions into something observable and repeatable, allowing me to critique the system’s strengths and pinpoint where it went wrong with defaulted or generic information.

Persona 1

Family first-time traveller.

Click to expand

Dewi is a 36-year-old primary school teacher from Jakarta taking her first overseas trip with her husband and two young kids to Tokyo. She values clarity, reassurance, and practical, family-oriented advice for her December winter holiday.

{
  "User Identity": "Dewi Santoso",
  "Trip Info": "Tokyo, Japan — from Jakarta, Indonesia — December — 7 nights (Dec 20–27)",
  "Demographics": "36 — Primary school teacher — Leisure",
  "Experience Level": "First overseas trip",
  "Interests": "Theme parks, easy sightseeing, local mild food, shopping malls",
  "Preferences": "Moderate — guided, family-friendly, simple planning",
  "Companions": "Family of four (two young children)"
}

output

Persona 2

Independent design-savvy solo explorer.

Click to expand

Jordan is a 28-year-old UX designer from Sydney spending 12 days in Kyoto in April. A confident and frequent traveller, Jordan values insider knowledge, aesthetic experiences, and efficient logistics over beginner guidance.

{  
	"User Identity": "Jordan Lee",  
    "Trip Info": "Kyoto, Japan — from Sydney, Australia — April — 12 nights",  
    "Demographics": "28 — UX designer — Leisure / Cultural exploration",  
    "Experience Level": "Frequent traveller",  
    "Interests": "Local art, design, architecture, coffee culture, photography, historical walks", 
    "Preferences": "Independent — mid-to-high budget, flexible schedule",  
    "Companions": "Solo traveller"
}

output

Persona 3

Comfort-first retired couple.

Click to expand

Harold (68) and Margaret (65) are retired teachers from Melbourne revisiting Europe for the first time in 15 years. They’re travelling to Paris for nine nights in September and want comfort, safety, and guided experiences without tech complexity.

{
	"User Identity": "Harold & Margaret Thompson",
	"Trip Info": "Paris, France — from Melbourne, Australia — September — 9 nights",
	"Demographics": "68 & 65 — Retired teachers — Leisure / Cultural discovery",
	"Experience Level": "Moderate (Europe once, 15 years ago)",
	"Interests": "Art museums, walking tours, French cuisine, gardens, scenic viewpoints",
	"Preferences": "Luxury — guided, minimal digital tools",
	"Companions": "Couple"
}

output

Visa info
Explicitly calls out Indonesian visa requirements and Visit Japan Web; this directly targets her passport and first-time status.
Family logistics
‍Mentions Airport Limousine Bus as easiest with kids, stroller hire + lockers, and child IC cards with half fares — therefore a strong family calibration.
Winter prep
Gives exact temps, sunset time, and layered clothing guidance; fully aligned with “tropical to winter” experience gap.
Food Comfort
Suggests mild spice staples (ramen, curry, tempura, gyoza) and adds Halal Gourmet Japan; a comfort and farmiliar recommendation.
Connectivity
Covers eSIM/pocket Wi-Fi, Type A/100V, and recommends comprehensive travel insurance — all high-utility and persona-specific.

Insider picks
Delivers Ando, Miho Museum, KYOCERA, MoMAK, specific coffee spots; strong alignment with design/UX/photography angle.
Seasonal timing for April (cherry blossoms) Calls out April conditions, tail-end blossoms, blue hour/sunrise slots; clearly tuned to timing + photography.
Streamlined itinerary logic
Smart plan (linking close neighbourhood; and early starts for busy attractions), close but could be more planned and precise.
Advanced tips and information
Covers crowd timing, restrictions, reservations; but is missing deeper hacks, so it’s solid but not maxed-out for seasoned travellers.

Simplest arrival plan Effective: pushes official taxis, de-prioritises train for comfort
Connectivity & Payment
Basic and reassuring. Simple explanation of tipping, cards, cash, adaptors, and travel eSIM; accessible and calm.
Confort
Emphasis on car over train, jet lag easing, mild walking, queues avoided — strong comfort logic.
Activities and rest
Timed entries and museum closures are covered, but “guided tours” and explicit rest breaks/accessibility are implied rather than clearly recommended.
Accessibility cues
Covers accessibility for public transport, but lacks detailed accessibility cues for each activity/location.

Omit Unnecessary
No unnecessary noise info appears, such as JR Pass, nightlife, coworking, deep dives. Scope discipline is good and keeps cognitive load low.
Reassuring and structured tone
Reads like a calm checklist for nervous parents, not a exciting hype — exactly what parents need.

Includes airport train vs bus breakdown, Visit Japan Web, IC card setup, ATM advice — too much onboarding for a “frequent” traveller; this dilutes the persona fit.
NO Generic top-10s
No kid fluff, no shallow top-10 spam; recommendations are curated and on-theme.
Few but high-value insights
Strong curation on spots and timing, but space is spent on things Jordan likely knows already, so some optimisation is needed.
Prioritises discovery, not reassurance
Good on discovery, but reassurance/logistics creep in — the program is caught between “experienced” and “semi-new”. It only partially trusts Jordan.

Omit
No nightlife, no budget gaming, no app maze; tone stays age-appropriate.
Calm, trust-building tone
Language is steady, direct, and respectful

It works!

Across the three personas, the AI demonstrates credible personalisation range, from Dewi’s guided reassurance, through Jordan’s creative autonomy, to Harold and Margaret’s comfort-centric clarity. The system clearly adapts to intent, region, and confidence level.

These persona evaluations hint where future iterations should cut redundancy and deepen contextual sensitivity. While proof that the personalisation is working, it still needs further refinement.

how does it work?

(Basically)

[Backend]

01 Gathering Known Data

The program begins by combining any existing user information—such as travel preferences, past destinations and experiences, and demographic details such as occupation, age and gender — with real-time data about the searched destination, including weather, hotels, activities, customs requirements and cultural factors.

James was here

[Backend]

1. Webflow: Collecting and Sending User Data

‍2. Vercel: The Middle Layer

‍Vercel acts as the server-side logic layer. It receives the JSON payload from Webflow and forwards it to the Wordware API.

02 Persona Expansion

Using the combined input, the system builds a more detailed travel persona. This includes generalising traits and expanding on inferred needs, like budget considerations, temporal experience, language support, climatic needs...

James was here

[Backend]

3. Wordware + ChatGPT: AI Personalisation Pipeline

‍This is where the real magic happens. Wordware runs a three-step internal AI process (using ChatGPT ) to generate hyper-personalised output:

Persona Building: It expands on the input data to create a rich persona. For example, if a user says they're from Indonesia, it infers relevant contextual traits (e.g., culturally diverse, tropical climate, language background).
Gap Identification: It analyses what the user might not know about the destination — identifying gaps between their background and where they’re going. E.g., someone from Indonesia visiting Japan might need info on colder weather, cultural norms, or local etiquette.
‍Content Generation: It produces well-written, friendly content that feels human and relevant, delivered in structured JSON (headings, paragraphs, etc.).

03 Personalised Checkout Page

Finally, the program generates a customised post-search checkout page. This page includes tailored recommendations for accommodations, activities, health and safety tips, required travel documents, and local transport options—curated to align with the user's persona and destination data.

James was here

[Backend]

4. Back to Webflow: Displaying the Output

‍The AI’s final output is sent back to Vercel, which filters and returns the cleaned JSON. Webflow’s JavaScript then takes that response and dynamically inserts the content into the appropriate divs on the page — no page reloads, no buttons, just seamless personalisation.

The story:

Every project starts with curiosity, and this one began with a question: What if AI could make a website feel genuinely personal—without the user ever typing, clicking, or asking?

Phase 01
Start

Assumptions

This project began as a university brief to apply AI meaningfully—find a problem space and design an intervention. I started fast: a Wordware.ai prototype that ingested a few simple inputs (name, age, nationality) and chained LLM steps to infer personality traits and recommend accommodation. It worked—but that early success exposed a bigger truth: LLMs aren’t recommendation engines. They’re language models. What I’d built competed with mature, data-driven algorithms already used by Booking.com, Expedia, and Trip.com, and brought no clear advantage. That realisation was a turning point: learn the difference, and use the right tool for the job.

Rather than force LLMs into matching tasks, I reframed the problem around their natural strength—language and context. If algorithms are excellent at ranking hotels, what’s missing is the human layer: the “why,” the nuance, the for-you context that lives in copy, guidance, and tone. I shifted the focus to the post-purchase moment (checkout/confirmation), where personalisation usually disappears. The question became: What if AI could quietly reshape page content based on known user data—no chat prompts, no buttons—so the experience felt specific, timely, and helpful?

Concept Pitch Deck

Phase 02
Concept

Refined Concept

From there, the concept crystallised: expand the user profile with contextual inference, identify gaps between a traveller’s background and destination (e.g., a warm-climate traveller heading to a cold country; a family vs a solo traveller), and generate dynamic, structured content—packing advice, etiquette notes, seasonal tips, transport options—inserted directly into the page.

Phase 03
MVP

Iteration

I iterated the AI pipeline extensively in Wordware: refining prompt stages (Persona → Gap Identification → Output), enforcing structured JSON, and tuning style and tone. Along the way, I wrestled with three core LLM challenges:

Useful depth vs fluff: securing concrete, destination-specific advice rather than generic filler. (Improved with prompt restructuring and model upgrades—from early runs to later models that handled reasoning better.)
Avoiding stereotypes: removing brittle assumptions tied to nationality/occupation; adding guardrails and neutral, evidence-based phrasing.
Structure vs creativity: keeping JSON predictable for the UI while leaving room for delightful, serendipitous suggestions (nearby cities, seasonal events).

Phase 4
Prototype Development

Low Fidelity Prototype

Mid-sprint I built a working demo with a local site: a Python script gathered inputs, called Wordware, and returned clean text. I briefly integrated a hotel search API to add live properties and used AI to explain why a place might suit a given traveller—but I cut it due to token costs and because it risked drifting back into “recommendation engine” territory. The learning: be ruthless about scope; keep the AI where it adds unique value.

Our Slide Deck

Webflow project

Github code

Phase 5
Prototype Refinement

Personal Project

After the class and semester ended, I've kept going. I rewrote the integration from scratch: Webflow on the front, Vercel as the server layer, Wordware for the AI pipeline. Webflow JS now collects inputs, Vercel orchestrates the request, Wordware returns structured JSON (headings, paragraphs, cards), and the UI injects content directly into the right components—no reloads, no prompts. The result is a proof-of-concept that feels like the page simply knows you.

What I learned (and applied):

Product judgement: don’t chase novelty; place AI where it’s comparatively strong (language, explanation, tone, context).
Invisible AI is a design decision: proactive, context-aware content beats another chatbot.
Technical depth: API design, JSON contracts, async flows, prompt design/guardrails, and front-end injection patterns for dynamic content.
Data viability > data hunger: use the rich data travel companies already have; don’t burden the user.
Ethics & inclusion: steer clear of cultural stereotyping; prioritise neutral, practical guidance.I’m genuinely energised by where this landed. It’s not a chatbot and it’s not a recommender—it’s a quiet layer that makes the web feel more considerate. And personally, this project pushed me from “I’ve never wired an API” to shipping a functioning AI-backed experience—end-to-end thinking, careful scoping, and a build that proves the point.

🚀 Future
Directions

Performance and Speed

Right now, the system’s biggest limitation is speed. Each round of reasoning inside Wordware adds latency, and Wordflow itself slows down the response chain. In future iterations, I plan to interface directly with the ChatGPT API rather than relying on Wordware, allowing tighter control over token usage and request flow. By streamlining the logic and reducing token count, the system could return results faster while still maintaining contextual depth.

Adapting to New Models

This project began on ChatGPT-3.0 and later upgraded through 4.0 and 5.0, with each model bringing exponential improvements in reasoning, context retention, and nuance. As these models evolve, the prompting strategy will need to evolve too. Where earlier versions required explicit instructions to “consider culture, weather, and group type,” newer ones infer these automatically. The design challenge ahead lies in re-tuning prompts and logic to match each model’s strengths, ensuring the system remains efficient, adaptive, and intelligent.

AI

+

Design

The idea:

AI That Quietly Shapes the Experience

Not Your Typical UX Project

This is a long question or heading to teaser the hidden content

This is a long question or heading to teaser the hidden content

This is a long question or heading to teaser the hidden content

Why travel websites?

Problem Area

Defining the Gap in Post-Purchase Personalisation

Research

Research Overview: Understanding what the landscape is like (data, user journey, and expectations).

Research Insights

Existing Solutuons

Existing solutions:

Opportunity Areas

Competitor breakdown

This is a long question or heading to teaser the hidden content

Where My Concept Diverges

Core Design Insights

the product

Heading

Your Tokyo family holiday is confirmed!

Smooth arrival and entry

Winter-ready packing and family days out

Getting around, money and easy eats

Hidden Logic

Your Kyoto spring getaway is confirmed, Jordan!

Touchdown to town, the smooth way

April rhythm: design, light and breathing space

Pack smart, move thoughtfully

Hidden Logic

Your Paris escape is confirmed, Harold & Margaret!

Arriving with ease

Art, gardens and viewpoints — without the queues

Dining and daily practicalities, tailored to you

Hidden Logic

The goal

Output / Outcome

Persona 1

Persona 2

Persona 3

It works!

how does it work?

(Basically)

[Backend]

01

Gathering Known Data

1. Webflow: Collecting and Sending User Data

‍2. Vercel: The Middle Layer

02

Persona Expansion

3. Wordware + ChatGPT: AI Personalisation Pipeline

03

Personalised Checkout Page

4. Back to Webflow: Displaying the Output

The story:

Research Overview:
Understanding what the landscape is like (data, user journey, and expectations).