All projects
Otto the Agent
As the sole product and UX designer for Otto, an AI-powered travel agent for business travelers, I led the design of a conversational interface that made trip planning feel as natural as messaging a human assistant. Working closely with the Founder/CEO, CTO, and a small development team, I helped shape the product from concept to MVP. Otto learns a traveler’s preferences almost instantly and proactively plans upcoming business trips by integrating with their Calendar. I focused on crafting interactions that felt highly personalized, efficient, and trustworthy, balancing automation with just the right amount of user oversight. My work spanned user research, interaction design, and prototyping, and directly contributed to Otto’s early success with enterprise users.
Company
Otto
Project length
9 months
Role
Product & UX design
Tools
Figma
Problem
Business travel is time-consuming, repetitive, and often stressful, especially for frequent travelers juggling packed schedules. Traditional corporate travel tools tend to be rigid, outdated, and optimized for compliance rather than user experience, forcing travelers to sift through generic results, re-enter the same preferences, and manually cross-check schedules.
One of the most significant pain points we uncovered was the lack of personalization and proactive support. Travelers get stuck reacting to plans rather than receiving intelligent suggestions tailored to their needs. And when things go wrong during a trip, like a flight cancellation, they often have to make phone calls or go through difficult workflows to fix the problem. EAs and travel agents are great at solving these problems for travelers, but most business travlers don't have that luxury.
Despite advances in AI, few tools leveraged user data like calendar context or historical behavior in meaningful ways. This gap left travelers overwhelmed by choices and under-supported, ultimately reducing efficiency and satisfaction in a process that should feel effortless. Frequent travelers often take the same trip to a particular office or meeting location, and they often just want to rebook the same flights and hotels, but there isn't an easy way to do that.
One of the most significant pain points we uncovered was the lack of personalization and proactive support. Travelers get stuck reacting to plans rather than receiving intelligent suggestions tailored to their needs. And when things go wrong during a trip, like a flight cancellation, they often have to make phone calls or go through difficult workflows to fix the problem. EAs and travel agents are great at solving these problems for travelers, but most business travlers don't have that luxury.
Despite advances in AI, few tools leveraged user data like calendar context or historical behavior in meaningful ways. This gap left travelers overwhelmed by choices and under-supported, ultimately reducing efficiency and satisfaction in a process that should feel effortless. Frequent travelers often take the same trip to a particular office or meeting location, and they often just want to rebook the same flights and hotels, but there isn't an easy way to do that.
Goals
Our primary goal with Otto was to redefine how business travel is planned and managed. We shifted away from rigid forms and static dashboards to a conversational, AI-first experience that feels like working with a trusted assistant. We aimed to design an interface that could rapidly learn individual traveler preferences, proactively plan trips by syncing with Google Calendar, and solve problems mid-trip with minimal user input.
Personalization, trust, and control were key pillars. Users needed to feel confident in Otto’s recommendations while retaining the ability to edit or override decisions at any step. The experience had to be fast, intuitive, and context-aware.
As a pre-seed startup, Otto also had a parallel goal to validate the product vision and prove it could be executed. In addition to winning over business travelers, we needed to build a compelling MVP that demonstrated technical feasibility and user value. A strong, well-designed product was critical, not just to attract early adopters, but also to get top-tier investors excited.
Personalization, trust, and control were key pillars. Users needed to feel confident in Otto’s recommendations while retaining the ability to edit or override decisions at any step. The experience had to be fast, intuitive, and context-aware.
As a pre-seed startup, Otto also had a parallel goal to validate the product vision and prove it could be executed. In addition to winning over business travelers, we needed to build a compelling MVP that demonstrated technical feasibility and user value. A strong, well-designed product was critical, not just to attract early adopters, but also to get top-tier investors excited.
My role
I was the sole Product & UX Designer for Otto, leading the design efforts from concept to MVP. I helped define the product vision and created the UX strategy for this AI-powered travel assistant. I was responsible for everything from user research and prototyping to final UI design, ensuring that the product’s conversational interface and personalized travel features met both business needs and user expectations. Additionally, I crafted product principles for Otto’s AI, which guided everything from tone and messaging to workflow paths, ensuring consistency and alignment with user expectations. Additionally, I crafted UX principles for Otto’s AI, which guided everything from the agent's tone and messaging to the workflow paths it followed.
A key achievement was designing the account profile and payments integration, enabling Otto to complete real flight and hotel transactions through its AI, a world’s first for an AI travel agent.
In addition to hands-on design work, I played a central role in aligning the product vision with technical feasibility, collaborating closely with engineering to integrate with Google Calendar for a high-impact traveler onboarding experience. My contributions were instrumental in driving Otto's early traction and setting the stage for a successful seed round fundraise.
A key achievement was designing the account profile and payments integration, enabling Otto to complete real flight and hotel transactions through its AI, a world’s first for an AI travel agent.
In addition to hands-on design work, I played a central role in aligning the product vision with technical feasibility, collaborating closely with engineering to integrate with Google Calendar for a high-impact traveler onboarding experience. My contributions were instrumental in driving Otto's early traction and setting the stage for a successful seed round fundraise.
Research & concept validation
We conducted several dozen interviews with frequent business travelers to uncover pain points and test early product hypotheses. Users consistently described travel planning as a fragmented and repetitive experience with outdated tools. They wanted a solution that could anticipate their needs and adapt to mid-trip changes with minimal effort. Users lit up when we asked them how they felt about a Otto being able to automatically get them home on an earlier flight if the last meeting of a trip is cancelled. "Magic."
Before building anything, we used Figma mockups and clickable prototypes to test a range of concepts and travel scenarios. We explored delivering the UI through different platforms, including Slack and SMS, because of their familiarity, daily usage, and strong notification systems. But we decided that these channels were likely too constrained for visually rich tasks like selecting a hotel, or complex workflows. This led us to design a dedicated web interface that could support more dynamic, high-context interactions. *Note that Otto's roadmap includes integrating these other platforms and modalities in the future, aiming to meet users in their typical flow of work.
We learned that users have almost no tolerance for inaccuracy with a product like Otto. If pricing isn't spot on, or we present hotel and flight options that aren't relevant, the whole experience falls apart. So we had a high bar to meet.
Before building anything, we used Figma mockups and clickable prototypes to test a range of concepts and travel scenarios. We explored delivering the UI through different platforms, including Slack and SMS, because of their familiarity, daily usage, and strong notification systems. But we decided that these channels were likely too constrained for visually rich tasks like selecting a hotel, or complex workflows. This led us to design a dedicated web interface that could support more dynamic, high-context interactions. *Note that Otto's roadmap includes integrating these other platforms and modalities in the future, aiming to meet users in their typical flow of work.
We learned that users have almost no tolerance for inaccuracy with a product like Otto. If pricing isn't spot on, or we present hotel and flight options that aren't relevant, the whole experience falls apart. So we had a high bar to meet.
Test & iterate
We quickly moved past Figma mockups to explore core conversation flows and travel planning logic with the AI agent. We built a working prototype using Chainlit, allowing us to test real-time conversations with the AI in a lightweight UI. This let us evaluate how the agent performed in live scenarios like planning trips or understanding a user's travel preferences based on historical calendar data. This tight design–build–test loop helped us fine-tune both the interaction model and overall experience.
Of course, working with a conversational AI brought some challenges. Early versions of the assistant struggled with brittle conversation flows and often got confused. We also saw hallucinations, including one case where the agent confidently booked a hotel that didn’t exist, resulting in a real credit card charge. These issues forced us to rethink how we validate responses, control the AI’s output, and give users the right guardrails to catch and correct mistakes before they became costly. It was a constant balance between flexibility and reliability, and each test cycle helped us make the product smarter more trustworthy.
Of course, working with a conversational AI brought some challenges. Early versions of the assistant struggled with brittle conversation flows and often got confused. We also saw hallucinations, including one case where the agent confidently booked a hotel that didn’t exist, resulting in a real credit card charge. These issues forced us to rethink how we validate responses, control the AI’s output, and give users the right guardrails to catch and correct mistakes before they became costly. It was a constant balance between flexibility and reliability, and each test cycle helped us make the product smarter more trustworthy.
Outcome
The launch of Otto’s early product delivered meaningful traction across both user experience and business validation. Business travelers who tested the product responded positively to the AI assistant’s ability to quickly learn their personal preferences, plan trips proactively, and reduce the mental overhead of managing travel. Users especially appreciated how Otto maximized trip quality within company policy, helping them make the most of each trip while staying fully compliant. This level of intelligent support made the experience feel both efficient and personalized which led to higher trust and satisfaction.
We launched to an alpha waitlist of thousands, demonstrating that it was technically feasible to plan and book trips with real transactions entirely through AI. This included the industry-first milestone of completed bookings autonomously. The combination of strong user enthusiasm and proven execution helped Otto raise a $6 million seed round led by Madrona and Direct Travel. They are now building out a beta product.
We launched to an alpha waitlist of thousands, demonstrating that it was technically feasible to plan and book trips with real transactions entirely through AI. This included the industry-first milestone of completed bookings autonomously. The combination of strong user enthusiasm and proven execution helped Otto raise a $6 million seed round led by Madrona and Direct Travel. They are now building out a beta product.
