"Trip planning shouldn't feel like a second job. That was the first thing I wrote on my FigJam board when I started this project and honestly, that one sentence drove every single design decision after it."
est read time 15 min
- What it is: Anna.ai is an AI-powered travel planning app prototype I built end-to-end to simplify trip planning.
- Problem: Trip planning today forces users to jump across multiple apps for discovery, pricing, booking, group coordination, and expense splitting.
- Research done:
- Market research
- Competitive analysis
- On-campus user interviews
- PM deliverables created:
- Personas
- Journey maps
- Business Model Canvas
- MVP scope + prioritized user flows
- What I built: A hi-fi Figma prototype (12+ screens) with a conversation-first AI assistant and feature modules for:
- Flight deal-finding agents
- Itinerary planning + sharing
- Group coordination
- Automated expense splitting
- Outcome: Designed to reduce cognitive load, minimize context switching, and help users go from idea → booked trip → synced group with fewer steps.
Anna.ai is an AI-powered mobile travel companion I designed from scratch: from the first sticky note on a FigJam board all the way to a high-fidelity Figma prototype.
The core idea is simple: right now, planning a trip is exhausting. You're jumping between Google Flights, Reddit, TripAdvisor, a group chat with your friends, and three different booking apps, all just to figure out where to go and how much it'll cost. Nobody has built one app that brings all of that together in a way that actually feels lightweight and easy to use.
That's the gap Anna.ai fills. One conversational AI that helps you find deals, plan your itinerary, split expenses with your group, and get travel inspiration but without the overwhelm.
In this project I am covering: Market Research → Competitive Analysis → User Interviews → Persona Building → Journey Mapping → Business Model Canvas → Lo-Fi Wireframes → Hi-Fi Prototype
Before I designed anything, I needed to understand why trip planning is broken. Not assume but by actually understanding it.
When I looked at the current experience, the friction points were everywhere. Users are spending hours across multiple platforms with zero continuity between them. There is no single place where you can discover a destination, check if flights are affordable, coordinate with your travel group, and split the final bill. Every one of those tasks lives in a different app.
And here is the thing about friction which I understood over the years in technology innovation and entrepreneurship, in product thinking, we know that every extra step between a user and their goal is a drop-off risk. Every unnecessary decision costs cognitive load. The travel planning experience today is basically one long chain of unnecessary decisions that nobody has bothered to simplify.
So the product question I started with was: what is the minimum number of interactions a user needs to go from "I want to go somewhere" to "my trip is booked and my group is synced"?
Anna.ai is my answer to that question.
I want to be clear about something, I did not start designing screens first. I know that is tempting, especially when you already have ideas, but I have learned that the most dangerous thing a designer or PM can do is fall in love with their solution before they have validated the problem. So I spent a good chunk of this project just researching.
The first thing I wanted to know was: is this even a real market opportunity, or is it just a personal frustration?
Turns out, it is a massive opportunity.
The AI in Travel and Tourism market is sitting at $2.95 billion in 2024 and is projected to hit $13.38 billion by 2030. That is not slow growth. That is a 4.5x increase in 6 years. The market signal is loud and clear which is that travelers want AI assistance, and the current solutions are not fully delivering yet.
A few more data points that shaped my thinking:
- 60% of travelers are already open to using AI for trip planning
- 40% of what I call "experimenter" travelers are actively testing AI tools right now
- 20% of millennials already rely on AI-generated itineraries
And when I looked at where AI is being applied in travel, there are three main areas: itinerary generators, chatbots, and dynamic pricing. The interesting thing is that nobody has connected all three into one coherent user experience yet and I saw this as an big opportunity.
Research tools I used: Statista, Google Search, Reddit, Google Gemini
I analyzed six competitors. I did not just look at features, I looked at their strategic positioning, what user segment they were going after, and most importantly, where each of them was leaving value on the table.
| Competitor | Their Angle | What They're Missing |
|---|---|---|
| Mindtrip | Socially-driven AI — blends group chat with shared itineraries | No automated deal-finding, no expense splitting |
| Layla | Instant AI itineraries, handles complex multi-stop routing | Limited social/community features, fewer real traveler photos |
| Pilot | "Notion for trips" — collaborative planning with PDF export | Heavy for simple trips, no coupon or deal finder built in |
| Skoot | Family-first, kid-safe travel booking with destination comparisons | Very narrow audience appeal, deals interface not mature |
| Booked AI | Automated bargain hunting with a conversational agent feel | Zero community or inspiration feed |
| GuideGeek | Zero-install WhatsApp/Instagram chatbot, great for quick local tips | Not a full itinerary builder, no cost-splitting or group tools |
After mapping all of this, I created a Similarities, Differences, Learnings, and Opportunities framework to synthesize the findings. This is something I think is really underrated in product work, most people do a competitive analysis and then move on. But the real gold is in the synthesis of: what patterns do you see, what can you learn from what everyone is doing wrong, and where is the white space?
What I found everyone is doing the same: All of them have some version of AI chat as the core UI, some form of itinerary generation, and some level of booking or deal integration.
Where they are all different: Audience focus, platform scope, and how deep the deal-finding actually goes.
Key Learnings: Users love AI for research but hate cluttered UIs. They want to share trip plans with the people they travel with. And they strongly prefer lightweight apps that are fast and simple over power-packed platforms that feel like work.
Opportunities I identified: Mood-based trip filters (Beach Vibes, Roadtrip, Casino mode, etc.), a coupon finder built directly into the AI flow, a WhatsApp chatbot extension, real-time expense tracking, and even an AR destination preview for places you're considering.
Honestly, this was my favorite part of the whole project. I went on-campus at ASU and conducted real interviews with people who actually travel eg, students, professors, families, and international students. I wanted to learn the real stories about real trips that went well and trips that didn't.
I designed 17 open-ended interview questions specifically to avoid leading the participant toward any particular answer. I learned early on that if you ask "do you find trip planning stressful?" you are going to get "yes" because that is a leading question. It feeds them the answer. Instead, I tailored my questions in their past behavior rather than hypothetical future promises. I asked things like "tell me about a trip that didn't go well, what happened?" because users are highly reliable historians of their own past pain, but terrible at predicting what they might do in the future.
During the sessions, I leaned into core UX techniques to get bottom of the ocean full of opportunities to solve these painpoints. I learned to embrace the "awkward pause", simply waiting a few seconds after a participant answered. People naturally want to fill the silence, and that is usually when they dropped their guard and revealed deeper emotional frustrations. When they mentioned a specific pain point, I used the "Five Whys" technique to drill down to the root cause, rather than just accepting the first thing they said.
One of the most valuable things I looked for during these interviews were "workarounds." If someone told me they had to use three different apps, take a screenshot, and text it to their partner/group of friends just to coordinate a flight, I paid close attention. Whenever a user uses their own manual solution, it means the problem is painful enough that they are actively spending time trying to solve it. Those are the massive indicators pointing to real product opportunities.
After wrapping up the interviews, I had ton of raw qualitative data. To make sense of it all and turn it into actionable insights, I used affinity mapping. I pulled out the most impactful quotes and observations, grouped them by theme, and looked for the overlaps. I wasn't just looking for what one person said but I was looking for the patterns across all the users. Those are the insights the one's which will connect the dots.
From the interviews, I built 6 detailed user personas and I want to be clear that these are not made-up characters. These are AI-summarized versions of real people I actually spoke to on campus, yes we should use AI wherever necessary.
Here is a quick look at who they are and what they each need:
| Persona | Who They Are | Their Core Pain Point | What Would Help Them |
|---|---|---|---|
| Jordan | Geography undergrad, travels solo or with 2-3 friends | Wants student discounts but finding them takes too long | Study-aligned itineraries, automatic student deal highlights |
| Steve Choe | Professor, extremely budget-conscious, avoids upsells | Gets overwhelmed by cluttered booking flows and abandons | Genuine bargains shown upfront, option to book by phone |
| Meenakshi | Mom of 2, family road trips | Managing 4 schedules, finding kid-safe activities on the road | Family activity filters, mapped driving routes, calendar sync |
| Madhav | International student, Vegas trips | Live USD-INR exchange rate matters every single decision | Currency conversion built into deal displays, loyalty point tracking |
| Palak | Travels with husband, books through agents | Needs "husband-approved" options, prefers agents for complex trips | Couple-friendly package filters, offline-printable itineraries |
| Ankush | Solo traveler, best friend trips, last-minute booker | Can't stand upsells or complicated checkout flows | Clean no-noise UI, sub-2-tap booking, last-minute solo deals |
- 🎨 Figjam Board: https://www.figma.com/board/T4swU311ILU89s382PDeYz/Anna-AI--Market-Validation-?node-id=0-1&t=y5mPYEF1T04DfQZK-1
I didn't stop at interviews. I built User Journey Maps for 16 additional personas, including Kristen, Triveni, Diane, Manoj, etc to visualize their end-to-end travel experience. I mapped their paths across five key stages: Awareness, Consideration, Decision, Experience, and Post-Trip. For every single stage, I documented exactly what they were doing, what they were feeling, and where the specific pain points lived.
This is something I feel strongly about: journey mapping is not just a deliverable but they are more of a thinking tool. The goal isn't just to make a pretty diagram for a presentation. The real point is to find the "divergence points" which are those specific moments where a user’s expectations and their actual reality move apart.
When I looked at these maps, I wasn't just looking for a quick answer about what was wrong. I was looking for the makeshift fixes people created when the existing tools failed them. Those gaps between what a traveler expects and what they actually get are exactly where a product creates the most value. By identifying these friction points across 16 different types of travelers, I was able to pinpoint exactly where the most impactful opportunities for the product were hiding.
After all that research, I had a lot of information. The challenge now was: what do I actually build?
This is the phase that I think separates okay product thinking from great product thinking. It is easy to collect user feedback and build a feature list. It is hard to look at all that feedback and identify the underlying need that one elegant solution could address.
For Anna.ai, the underlying need was: reduce the number of context switches a traveler has to make between discovery, planning, booking, and coordination. Everything I designed flows from that.
Before I ever touched a design tool or sketched a single screen, I had to answer a hard question: Does this actually work as a business? I’ve seen plenty of great-looking products fail because they had no real path to staying alive. A good product without a viable business model is just an expensive side project.
To bridge the gap between a cool idea and a sustainable company, I built a full Business Model Canvas for Anna.ai.
I didn't just want a quick answer for how we’d make money; I wanted to understand the entire ecosystem. I looked at our Value Propositions to see how we specifically solve those manual fixes travelers are currently forced to do. I mapped out our Revenue Streams and Key Partners to ensure that the "divergence points" I found in my journey maps actually turned into opportunities for growth.
By starting with the business logic, I ensured that every design decision I made later on was rooted in creating value not just for the user, but for the long-term success of the product itself.
| Section | Anna.ai's Answer |
|---|---|
| Value Propositions | Personalized AI itinerary · Group chats & trip coordination · Expense splitting · Coupon finder · Flight deal finder · Real-time price alerts · Voice interface · AR destination preview |
| Customer Segments | Solo travelers · Families · Friend groups · Business travelers · Budget travelers · Luxury travelers · Teens · Explorers |
| Key Partnerships | Airlines · Hotel chains · Tourism boards · Local guides · Influencer platforms · API providers |
| Key Activities | Itinerary generation · Data analysis · UX design · AI training · Platform integration · Customer support |
| Key Resources | LLM · Travel data APIs · Dev + QA team · Cloud infrastructure · User insights · Knowledge base |
| Channels | Mobile app · Web app · WhatsApp bot · App stores · Social media · Ads · Email · API |
| Customer Relationships | 24/7 in-app support · Push notifications · Email · Loyalty rewards · Social travel forum |
| Revenue Streams | Freemium subscriptions · Advertising · Commission on bookings |
| Cost Structure | Cloud hosting · Supabase · Data licenses · R&D · Marketing · API fees · Dev team · Infra |
The revenue model I landed on was freemium + ads + commissions. Freemium lowers the barrier to entry for budget travelers (a huge segment, as my personas showed). Commissions on bookings create alignment between Anna.ai and the user: we both benefit when they find and book the best deal. And ads, if done right, can actually be useful in a travel context (think targeted local experiences or partner promotions in a destination you're already researching).
This is where all the research becomes something you can actually see and touch. I went through two rounds of design: lo-fidelity wireframes first to figure out the flows and structure, then hi-fidelity screens to make it feel like a real product.
I sketched three primary user flows before touching Figma. The reason I do lo-fi first is not because it is faster, it is because low-fidelity thinking forces you to focus on structure and flow rather than getting distracted by colors and typography. You cannot hide a bad UX behind nice visuals in a pencil sketch.
Flow 1 — Flight Deal Flow
The user enters from login, navigates to the Flight Deals section, selects an AI agent (Cheap Flight Agent, Coupon Finder, Track Price Drops, or Flight Comparison), gets results, and confirms a booking.
Flow 2 — Sharing Trip Plan Flow
User accesses their itinerary, taps the share option (triggered by voice or three-dot menu), selects contacts or a group trip, and confirms the share. This flow was specifically designed around one of my key research insights: travelers want to share plans with their travel companions without having to copy-paste or screenshot.
Flow 3 — Trip Expense Split Flow
User goes to My Trips, selects a trip, opens Trip Splits, sees the breakdown of who owes what, and either adds an expense or settles up. Clean, direct, no complexity.
This is the part I am most proud of. The hi-fi prototype has 12+ screens and covers the full user journey from first open to post-trip. Here is what each screen is doing and why it was designed that way:
Sign Up & Login Screen
I kept this as clean as possible just the essentials. Full name, email, password, and Google/Facebook SSO. The reason is simple, every extra field on a signup form is a drop-off risk. I wanted to minimize friction at the top of the funnel because acquiring a user who never completes onboarding is worse than not acquiring them at all.
AI Assistant + Slider Navigation
The main experience is conversation-first. Anna is an AI you talk to, not a form you fill out. The dark slide out navigation gives you access to Inspirations, My Chats, Trip Split, Flights Deals, Hotel Deals, and Search History but it lives behind a hamburger menu so it does not clutter the primary experience.
The design principle here is progressive disclosure: show the user what they need right now and hide complexity until they actively ask for it.
Inspirations Screen
This is a Pinterest-style feed of destinations and travel content. But the difference from a regular inspiration board is that each destination card has a contextual AI Pro Tip embedded in it, for example, "Visit Japanese friendship garden early in the morning to avoid the crowds."
I included this because my user research told me something interesting: travelers do not just want to know where to go, they want to feel like they are getting insider knowledge. That is the job of a good travel companion, and it is exactly what Anna should be doing.
My Trips Screen
A dashboard showing upcoming and past trips. Each trip has a View Plan and Edit button. This is the central hub for the user's travel life like past, present, future. The design prioritizes upcoming trips at the top with a clear status badge (Upcoming / Past) so the most relevant information is always first.
Trip Splits Screen
This is probably the most underrated feature in the app and one I am really happy with. You see your group members, the total balance, what you are owed versus what you owe, broken down by expense. "Dinner at Shibuya, you are owed $80.00" that level of clarity.
What I found in my research is that expense tracking in group trips is one of the biggest sources of friction and relationship tension. People avoid the conversation because it is awkward. Making it automated and transparent removes the awkwardness entirely. Good product design should dissolve social friction, not just functional friction.
My Chats Screen
Group and individual chat view, with three smart action buttons at the bottom: AI Summarizer, Create Expense, Create Poll. The AI Summarizer was a direct response to a user insight, when you are in a chaotic group travel chat, nobody wants to scroll back through 200 messages to find out if someone already confirmed the hotel. One tap and Anna summarizes the thread for you.
Flight Deals Screen
This is the deal-finding engine. You have four AI Travel Agents: Cheap Flight Agent, Coupon Finder, Track Price Drops, and Flight Comparison. Below that, you have Saved Flights and a Recently Used section.
The mental model I used here is exactly the same as a human travel agent, you do not call a travel agent and ask them to do everything at once, you call them for a specific job. The AI agents are designed around discrete jobs-to-be-done: find the cheapest option, apply a coupon, alert me when the price drops, compare carriers.
Search History & Profile
Search history is clearable and shows recent queries with enough context to understand what you were looking for without having to remember. Profile gives you Settings, Payment Methods, Travel Preferences, and Help, nothing excessive.
Hi-fi prototype demo
- Figma - Hi-fi prototype and screen design
- FigJam - The full research board: competitive analysis, user personas, journey maps, BMC, wireframes
- Statista + Google Search + Reddit - Market sizing and trend research
- Google Gemini - Secondary research synthesis
- On-campus ASU interviews - Primary qualitative user research
- Rapid Prototyping - Bolt.new
I want to be honest here, I started this project thinking I already knew what the app should do. I've planned enough trips to know the pain. But the research genuinely surprised me and changed the design. I did not know how differently a budget focused international student and a family traveler think about the same trip. I did not realize that the post trip memory sharing experience was something users cared so much about. I did not expect that several users would say the most important thing to them is just being able to trust the information they're getting.
Those insights came from the research. Not from me sitting at my desk making assumptions.
If there is one thing I would tell anyone building a product, spend more time in the problem space mostly knowing the customer pain points, mapping them down and try to solve the bottlenecks than you think you need to. The design phase is actually the easy part once you deeply understand who you are designing for and what they are trying to do.
Thank you for going through my project.