Interview with Jar Kuznecov, CEO, Wonderplan

This interview is with Jar Kuznecov, CEO, Wonderplan.

As CEO of Wonderplan, an AI travel planner, how do you introduce your work and the core problem in AI trip planning and hotel search you’re focused on right now?

At Wonderplan.ai, we’re revolutionizing how people plan their travels by leveraging AI to create truly personalized, day-by-day itineraries. Imagine moving beyond the endless tabs and fragmented research, straight to a perfectly crafted trip that feels like it was designed just for you.

The core problem we’re tackling right now in AI trip planning and hotel search is the pervasive issue of generic recommendations. Most AI tools today, while helpful, often churn out one-size-fits-all itineraries that fail to capture the essence of individual travel styles, budget constraints, trip duration, or even real-time availability of attractions. This creates a significant gap between the initial inspiration — that exciting moment you decide on a destination — and the overwhelming reality of actual booking and planning. Travelers get excited, but then face a fragmented process that quickly becomes a chore.

Furthermore, the hotel and accommodation search experience is still largely disconnected from the planning process. You might have a fantastic itinerary laid out, but then find yourself sifting through generic hotel lists, with no contextual understanding of how well a particular stay aligns with your planned activities. Recommending a hotel miles away from your daily adventures defeats the purpose of a well-planned trip. It’s a critical oversight that leads to wasted time and suboptimal experiences.

At Wonderplan.ai, we’re solving this by building a platform that doesn’t just generate itineraries, but truly understands the traveler. We’re focused on delivering context-aware, dynamic itineraries that integrate accommodation suggestions seamlessly — recommending hotels that are not only within your budget and style but are also strategically located to enhance your daily plans, ensuring a cohesive and delightful travel experience from start to finish.

What pivotal experiences or decisions led you to build and lead Wonderplan in the AI travel space?

For me, the journey behind Wonderplan started with a deep personal frustration. I’ve always loved the idea of travel — the discovery, the culture, the escape — but I absolutely hated the planning. I found myself constantly stuck in “tab hell,” juggling twenty different websites, cross-referencing reviews, and trying to piece together a coherent schedule. It was fragmented, overwhelming, and frankly, it often sucked the joy out of the trip before I even left my house.

The “aha” moment came when I saw how rapidly AI was evolving. I realized that we finally had the technology to solve this better than any human travel agent or static website ever could. I envisioned a future where planning wasn’t about manual labor, but about a conversation with an intelligent system that actually knew you — your style, your pace, and your budget. That conviction — that personalized, AI-driven itineraries were the inevitable future of travel — is what made me decide to go all-in.

Transitioning from that idea to leading a team has been a profound experience. There’s a real weight of responsibility when you realize thousands of people are relying on your code and your logic for their precious vacation time. We aren’t just building a tool; we’re responsible for the logistics of someone’s honeymoon or their first family trip abroad.

It’s incredibly exciting to be at the intersection of AI and travel right now. We’re watching two of the world’s most significant industries being transformed simultaneously. I truly believe we are at a unique turning point where the “work” of travel is finally disappearing, leaving only the experience. There’s never been a better or more necessary moment to be building in this space.

From analyzing over 600,000 itineraries, what counterintuitive traveler behavior most changed your approach to hotel ranking in search with a practical takeaway for hoteliers?

That’s a fantastic question, and it touches on one of the most profound insights we’ve gained from analyzing over 600,000 itineraries. What truly surprised us — and completely shifted our approach to hotel ranking — is that travelers don’t actually optimize for proximity to the single most famous landmark. Conventional wisdom suggests everyone wants to be right next to the Eiffel Tower or Times Square, but our data shows something much more nuanced.

We’ve identified what we call “base camp” logic. Travelers are actually optimizing for minimizing their total daily travel time across their entire itinerary, not just one point. This means a hotel that is slightly off-center but strategically located near a transit hub or a diverse cluster of their planned activities — like local food markets, parks, or specific cultural spots — often results in much higher satisfaction. In fact, we see that travelers who choose a 3 or 4-star hotel with a high “neighborhood fit score” report better trip experiences than those in a 5-star hotel that is geographically isolated from their actual interests.

The practical takeaway for hoteliers is clear: stop marketing yourself solely based on your distance to the #1 tourist attraction. Instead, market based on the type of traveler your specific neighborhood attracts and the unique cluster of local experiences around you. A hotel that highlights its proximity to a great coffee shop, a reliable metro line, and a local park is often more valuable to a modern traveler than one that just claims to be “near the monument.”

This insight is fundamentally changing how we think about hotel discovery. We are moving away from static search results toward a future where a hotel’s rank is dynamic and contextually tied to the specific logic of a traveler’s day-by-day plan. It’s about being the right home base for the right journey.

Before you ship an AI hotel recommendation or itinerary model, which guardrails or evaluation steps are non‑negotiable?

That’s a critical question. At Wonderplan, we operate under the understanding that in travel, a bad AI recommendation isn’t just a “bug” — it has real-world consequences. If an AI hallucination leads someone to a hotel that doesn’t exist or a closed attraction, that can ruin a vacation day. Because of that, our bar for shipping is significantly higher than in most consumer AI products.

We have four non-negotiable evaluation steps.

1. Factual accuracy. We run every recommendation against live data to ensure the hotel is open, the price is in the right ballpark, and the amenities are current. Trust is our most valuable currency, and hallucinations are trust-killers.
2. Geographic coherence. We’ve seen raw models recommend back-to-back activities that are two hours apart without accounting for transit time. Our system audits every itinerary to ensure it’s physically possible. If the logistics don’t make sense, the itinerary doesn’t ship.
3. Personalization fidelity. We test the model’s “listening” skills. If a user tells us they’re traveling with a toddler, and the model suggests a party hostel or a late-night bar crawl, that’s a failure. We use specific test suites to ensure the output actually respects the user’s constraints and style.
4. Diversity and bias audit. It’s easy for a model to default to the same ten “popular” properties because they have the most SEO weight. We proactively ensure we’re surfacing a diverse range of accommodations that genuinely fit the traveler — not just the ones with the best affiliate margins.

Ultimately, travel is a high-stakes emotional and financial investment. We don’t just ship “cool” AI; we ship reliable travel partners. Our commitment to these guardrails is what allows our users to actually book with confidence.

In balancing personalization and privacy for hotel search, which specific user data do you retain versus aggregate or delete to reduce risk while preserving insight?

That’s a profoundly important question, and it’s one we’ve wrestled with deeply. Balancing personalization with privacy isn’t just a technical challenge — it’s a core ethical commitment. Our approach is built around a clear framework: what we retain at the user level, what we aggregate anonymously, and what we simply delete to minimize risk.

What we retain at the user level, always with explicit consent, are the travel preferences expressed during a session — things like travel style, budget range, and group composition. These are the direct inputs that power the next personalized recommendation. We also retain saved itineraries and destinations, so users can seamlessly return and continue planning. If a user opts in, we may also retain booking history to further refine future personalization, but this is always transparent and controllable.

What we aggregate is anonymized behavioral data. This includes patterns like:

• Which types of hotels users click on versus skip,
• Which itinerary structures are frequently edited versus accepted, and
• Overall destination popularity trends.

We also look at satisfaction signals — for instance, if a user immediately re-plans after seeing a recommendation, that’s an aggregated signal of dissatisfaction that helps us improve the model. This data is stripped of any individual identity and used purely for statistical insight.

Crucially, what we delete includes precise location data beyond what’s immediately necessary for a search. We don’t need to know someone’s home address to recommend hotels in Paris. We also have defined retention windows for session data — if a user hasn’t returned, that data is purged. We actively work to prevent and delete any sensitive inferences — like health conditions or political views — that could theoretically be derived from travel patterns.

Our guiding principle is simple: build a system that gets smarter about travel patterns without becoming more invasive about people. The truly valuable insight lives in the aggregate; the significant risk lives in the individual record.

What single change in copy, labels, or UI most improved trust and conversion on AI-generated hotel results?

That’s a crucial learning for us. The single change that most dramatically improved trust and conversion on our AI-generated hotel results was deceptively simple: adding a short, specific “Why we picked this” explanation label directly under each recommendation.

Before this, users would see our AI-powered hotel suggestions — star ratings, price, photos — much like any traditional hotel search engine. While the underlying AI was doing incredible work to match hotels to their itinerary, the reasoning was opaque. Users often felt the recommendations were arbitrary, leading to skepticism. They frequently abandoned our platform to cross-reference on Booking.com or Google Hotels, eroding our conversion rates.

Introducing that simple label — for example, “Chosen because it’s 8 minutes from your Day 2 activities and fits your mid-range budget” or “Selected for its family-friendly amenities and proximity to the children’s museum in your itinerary” — was a game-changer. It immediately provided transparency. Users understood why the AI had made that specific recommendation for their specific trip. It transformed the recommendation from a black-box output into a thoughtful, personalized suggestion. Trust soared, abandonment rates dropped, and direct booking clicks increased significantly.

The deeper lesson is profound: people don’t inherently distrust AI recommendations because they’re AI — they distrust them when they feel arbitrary or unexplained. The moment you show your reasoning, you transform a black-box output into a trusted advisor. This principle of transparency in AI UX is now non-negotiable for us. It’s about making the AI’s intelligence visible and actionable — turning complex algorithms into clear, human-understandable value.

What technique do you use to embed long‑stay economics and cost‑of‑living into AI hotel recommendations for remote workers?

That’s a very timely question, as the remote worker segment is growing exponentially. For us, recommending hotels for remote workers is fundamentally different from recommending for a short-term tourist. A tourist optimizes for 3 nights; a remote worker optimizes for 30 to 90 days. The per-night price becomes almost irrelevant — what truly matters is the total monthly cost of living and working.

To address this, we’ve developed a “total monthly cost model” within our AI. This goes far beyond just the accommodation price, even factoring in potential long-stay discounts. We integrate data on average meal costs in that specific neighborhood, the availability and pricing of coworking day passes or monthly memberships nearby, local SIM card and data costs, and estimated daily transport expenses. This gives a much more holistic financial picture.

Beyond cost, we heavily weigh “productivity infrastructure.” For a remote worker, a cheap hotel with unreliable WiFi or a noisy environment isn’t cheap at all — it’s expensive in lost productivity. So our AI also factors in verified WiFi speed ratings, proximity to dedicated coworking spaces, and quiet environment scores. We then distill this into a clear “remote work score” surfaced alongside standard hotel metrics, so digital nomads can compare options at a glance.

The core insight is that for this segment, our AI has to think less like a hotel search engine and more like a relocation advisor. It’s not “where to sleep for a few nights” — it’s “where to live and work effectively for a month.” That shift in framing changes everything about how we build the recommendation model.

If a hotel wants to be “AI‑ready,” which operational data streams or guest touchpoints should they instrument first to perform well in AI-driven search and planning?

That’s a question we discuss frequently with our hotel partners. For a hotel to truly thrive in an AI-driven search and planning ecosystem, they need to shift their mindset from static marketing to dynamic data provision. The hotels that will win are not necessarily those with the biggest marketing budgets; they’re the ones with the cleanest, most structured, and most real-time data.

The first and most critical step is real-time availability and dynamic pricing via API. If an AI planner can’t confirm a room is bookable at a current price, that hotel simply won’t be recommended. Hotels without live feeds are effectively invisible to modern AI planning tools.

Second, structured amenity data is essential. It’s not enough to say “free WiFi.” AI needs specifics: what’s the average speed? Are there dedicated workstations? What are the pool hours? What’s the check-in flexibility? These granular details allow AI to match a hotel to a traveler’s precise needs — far beyond generic marketing copy.

Third, guest review sentiment at the attribute level is crucial. Beyond a star rating, AI learns from tagged feedback like “great for families,” “quiet neighborhood,” or “excellent breakfast.” This rich qualitative data is what our AI uses to understand a hotel’s true character and match it to specific traveler profiles.

Fourth, detailed neighborhood and proximity data should be provided. Hotels should offer structured information about what’s within walking distance, transit options, and nearby attractions. This is what allows AI to build context-aware recommendations — ensuring the hotel fits seamlessly into the traveler’s planned itinerary.

The overarching point is this: data quality is the new SEO. It’s about making your property intelligently discoverable, not just broadly visible. Data is the bridge between your rooms and the perfect guest.

For teams building AI travel products, what one investment this year will yield the biggest leap in itinerary quality or hotel relevance?

If I had to pick one investment that will yield the biggest leap in itinerary quality and hotel relevance this year, it’s unequivocally building a high-quality proprietary feedback loop.

Right now, most teams are obsessed with the generation side — better prompts, more sophisticated foundation models, and more data at the input stage. While those factors matter, the real compounding advantage comes from the evaluation and feedback side. You need to systematically capture what users actually do after they receive an AI recommendation. Did they accept the itinerary? Did they edit it significantly? Did they ignore a hotel suggestion, or — crucially — did they book it and come back for more?

A model that learns from a million real traveler decisions — their edits, their bookings, their rejections — is worth more than any amount of prompt engineering. This is especially true for hotel relevance. The gap between “AI thinks this is a good hotel for you” and “the traveler actually booked and loved this hotel” is where the real signal lives. The teams that pull ahead will be the ones that bridge that gap by making the feedback loop a core part of their infrastructure.

The principle is simple: in AI, the quality of your feedback loop is the quality of your product. If you aren’t learning from how users interact with your outputs, you’re just guessing. The teams that instrument this well this year will be two to three years ahead of everyone else by next year — because they’ll be building on a foundation of real-world validation that others simply can’t match.

Thanks for sharing your knowledge and expertise. Is there anything else you'd like to add?

Thank you so much for the conversation — it’s been a pleasure to share our journey.

If there’s one thing I’d like to leave you with, it’s that we are still in the very early innings of what AI can do for travel. Beyond the tech, we must remember that travel is one of the most meaningful things people do with their limited time and hard-earned money. Getting AI right in this space isn’t just about better algorithms — it actually matters for people’s lives. A well-planned trip can be a core memory for a family or a life-changing experience for a solo traveler.

My call to the rest of the industry is this: our opportunity isn’t to replace the joy of travel with automation. It’s to use AI to remove every ounce of friction, stress, and “tab hell” so that the joy is all that’s left. We want to take the “work” out of travel so people can get back to the “wonder” of it.

I’m incredibly grateful for the chance to talk about this mission today. At Wonderplan, we believe the future of travel isn’t just about where you go — but how effortlessly you get there.

Our goal is simple: make the planning as delightful as the trip itself.