AIBusinessStrategySmall Business

An AI Travel Assistant That Stays With You for the Whole Trip

By Armando J. Perez-Carreno / Featuring Vipul Doshi

I talked with Vipul Doshi, founder of Zenvoya, about building an AI travel assistant that plans around your preferences and keeps helping after you book, plus why industry AI needs guardrails.

Most travel sites sell you a flight or a hotel and then they're gone. If your flight is delayed at 11 PM, you're the one calling the hotel, waiting on hold, and hunting for a place to sleep. Vipul Doshi, founder of Zenvoya, spent 20 years in travel customer experience, and he's building an AI assistant that gives regular travelers the kind of help that used to belong only to executives with an assistant.

In this episode, I talked with Vipul Doshi, the founder of Zenvoya. Vipul started as a programmer in 1989 and later became the founding CEO of IGT Solutions, a customer experience company that served airlines, hotels, and travel agencies for more than 20 years. That company was built around people helping travelers. With Zenvoya, he's taking the same goal and asking whether AI can do it for everyone.

My family loves to travel, so this one was personal. Planning a trip usually means one site for flights, another for hotels, another for cars, and another for things to do. Vipul's first point was that most people don't travel alone. One person plans, and then everyone has an opinion. The kids want one thing, someone else wants museums, and friends all have different budgets. Zenvoya lets a group collaborate on the plan, and each person can still book their own pieces.

The search works differently too. On a typical booking site, you enter dates, airports, and a class, and you get a long list to filter. On Zenvoya, you describe what you want in plain language. You might want somewhere warm, vegetarian-friendly, a Marriott or Hilton, free parking, and an aisle seat, or else a later flight. The system sorts your must-haves from your nice-to-haves and comes back with a few options that fit. It also remembers. If you told it you're vegetarian on your last trip, you shouldn't have to say it again.

I had a recent example of why that context matters. At a hotel where I was giving a workshop, I asked an AI assistant to help me pick a room in the app. It turned out part of the building had been remodeled and part hadn't, and the assistant helped me land in the remodeled part, away from the elevator and the gym. On a family trip, one of us wanted a pool, and two of the hotels we liked had pools closed for renovation. Vipul pointed out that hotels don't always advertise that. Guests mention it in reviews, and Zenvoya reads those reviews before it recommends anything.

The part that stood out most is what happens after you book. Zenvoya's assistant, Zoya, watches your trip. If your flight is going to land late, it tells the hotel so they don't give away your room. If rain is coming on your park day, it suggests a show it knows you'd like and can buy the tickets. Vipul has sat at O'Hare while a flight got pushed back an hour at a time until it was canceled at 11 PM. Zoya watches for that pattern and offers to book a refundable room nearby before the cancellation happens.

For the business owners listening, his distribution strategy is worth noticing. Vipul looked at the balance sheets of the big online travel agencies and saw that roughly half their revenue goes to advertising, and travelers end up paying for that. Zenvoya is live in the US, but instead of trying to outspend them on ads, it offers its platform as a white-label product. An airline, a retailer, or a food delivery app that already has loyal customers can add travel under its own brand. Partnerships like that are a real alternative to buying traffic, especially when you don't yet know your return on ad spend.

We also talked about why general chat tools aren't enough for something like this. Vipul's point was simple: if the AI makes up a flight that doesn't exist, you're stuck at the airport. Tools built for one industry need guardrails, like only recommending flights that come from a real data source and checking the results before showing them. I mentioned Jev, a new decision model I've been testing, as one way to run a fast final check like that.

My takeaway is that the best uses of AI come from going deep into one industry and fixing the pain people already feel. Vipul's advice for travelers is to use the tools to plan more carefully, and to know ahead of time who will help you if something goes wrong at midnight in an unfamiliar place. You can try Zenvoya at zenvoya.ai.

Published by Armando J. Perez-Carreno

Work with us

Let's find the busywork eating your week.

Book a free discovery call. We look at how your team spends its time, point to the tasks AI can take off your plate, and treat you like family from the first hello.