Year One of Building: What We Are Shipping and Learning
A year into building Nowah: foundations, milestones we are aiming at, surprises from private testing, and the plan for year two — without pretending we already scaled.
One year ago, Nowah was an idea and an empty repository. Today it is a product in active build: private testing, live inventory integrations, and a trajectory we still have to earn in public. This is the honest timeline of the build — not a growth story we do not have yet.
Quarter one: foundation
The first three months were pure building. No users, no revenue, no external validation. Just code, architecture decisions, and the conviction that conversational AI could replace the broken search-and-browse model.
We built the core agent with its tool suite. We integrated with our travel data provider for live flight and hotel inventory. We implemented the streaming conversation architecture. We built the mobile app, the web app, and the shared backend.
The milestone of Q1 was the first successful end-to-end test: a natural language trip request, processed by the agent, resulting in a real booking against live inventory. It worked. Not perfectly, but it worked.
Quarter two: first users
We put the product in front of real people and everything changed. The bugs surfaced immediately. The surprising use patterns appeared. The gap between our assumptions and user reality became visible.
The milestone of Q2 was the first real booking. Real money. Real flight. Real user going to a real destination. It landed hard when it happened.
We also built streaming responses based on early user feedback. Showing the agent working in real time transformed the trust dynamic overnight. Users who had been skeptical became engaged once they could see the AI searching and reasoning.
Quarter three: growth
Word started spreading. Users who had great booking experiences told friends. The organic growth coefficient began climbing. We are shipping trip management, voice input, and significant agent intelligence improvements.
The milestone of Q3 was the first repeat booking, a user coming back for a second trip. This was more validating than any other metric because it proved the product was good enough to return to.
We also experienced our worst bug during this quarter, the double-charge incident that led to our three-layer booking safety system. Crisis compressed learning. The system is dramatically more robust because of it.
Quarter four: scale
The final quarter was about making everything work at scale. Performance optimization. Reliability hardening. Agent accuracy improvements based on thousands of conversations. Web app launch alongside the mobile app.
The milestone of Q4 was reaching a conversation and booking volume that made the aggregate data meaningful. We could now see real patterns in how travelers use an AI agent, and those patterns informed product decisions with statistical confidence.
The five biggest surprises
One: users anthropomorphize the agent immediately. They say please and thank you. They apologize for being indecisive. The agent's tone and personality matter far more than we expected.
Two: post-booking is as important as pre-booking. Users want the agent to help after the booking, not just during it. This shifted our roadmap significantly.
Three: the say-do gap. Users ask for cheap and choose convenient. Behavioral data contradicts stated preferences consistently.
Four: speed matters more than comprehensiveness. Users prefer a fast, focused recommendation over a slow, exhaustive analysis.
Five: the feature we were most proud of building had near-zero adoption. The feature we almost did not build became the most used.
Year two
Year two is about three things.
Proactive features: the agent starts suggesting opportunities before you search. Calendar awareness, price monitoring, and preference-based recommendations.
Platform API: the first endpoints that let other companies embed Nowah's intelligence into their products.
Expanded coverage: more destinations, more booking types, more post-booking capabilities.
The foundation is solid. The thesis is validated. Year two is where we build on it.
Nowah is an AI travel agent that searches and books real flights and hotels through conversation — no filters, no thirty open tabs. Plan your next trip.