They Built a Travel Company Few Understood. Now Buyers Are Paying Attention.

Editor's note: Sam and Devon are what I’d call “boomerang founders.” They built in this exact category before, sold during the pandemic when travel budgets collapsed and the market effectively froze, and then made the decision to come back.

It’s easy to forget what it was like trying to build a travel company when no one wanted to fund travel and no one wanted to buy travel software. Entire budgets disappeared overnight. Roadmaps became survival plans. Many founders left the category for good.

Sam and Devon didn’t.

What makes this conversation compelling is not just what Juno is building, but why they chose to rebuild it. This time, they’re returning with a decade of domain knowledge, clear founder-market fit and a fundamentally different toolkit in the AI era. They’re not just iterating on the same product. They’re rethinking coordination, pricing and experimentation in a world where software can compress decisions and align directly to outcomes.

There are real lessons here for buyers about how to evaluate vendors who return to a category with deeper domain knowledge, sharper pricing alignment and a more mature understanding of what actually breaks inside your workflows, and for entrepreneurs about conviction, timing and what it means to finish what you started.

As always, let us know what you think and please reach out if you have feedback.

- Dave

Enterprise travel teams know the problem: employee travel is well instrumented. Guest travel is not.

Candidates, interns, contractors, speakers and healthcare professionals still get routed through phone calls, spreadsheets and travel agents using infrastructure built decades ago. Travel managers own airline contracts and hotel negotiations, but for non-employees, coordination often lives in inboxes and manual workflows. And when guest travel can represent 10–20% of spend, or 50–80% in certain industries, that operational gap becomes expensive.

That is where Sam Felsenthal [LinkedIn] and Devon Tivona [LinkedIn] and Juno come in. Juno is a modern travel and expense platform purpose built for guests and non-employee travel. Rather than rip and replace core travel infrastructure like the GDS, Juno focuses on modernizing the interface layer, increasing employee efficiency while preserving the underlying contracts, integrations and duty-of-care systems enterprises rely on.

In the AI era, that coordination layer gets even more powerful. Juno now ties pricing to traveler satisfaction, moving from per-trip fees toward outcome alignment. This shift has real implications for procurement teams.

In this conversation, Sam and Devon cover:

• Why ripping and replacing 30-year-old infrastructure often causes more disruption than value
• How guest travel quietly becomes 50–80% of spend in industries like energy, education and field ops
• Why the most important layer to modernize is the interface, not the infrastructure
• How AI compresses 70–90 trip decisions into automated coordination
• Why Juno now surveys travelers and does not charge for poorly rated trips
• How small, high-agency teams experiment faster in the AI era
• Why founder-market fit matters more than product-market fit in niche verticals

If you care about vertical software, outcome-based pricing, AI-enabled coordination or building durable companies in legacy industries, this conversation explores what it really takes to modernize workflows without breaking the systems enterprises depend on.

Introductions

Dave Ambrose (01:01.76)

First off, thank you guys for doing this. It’s kind of surreal that we’re here kind of doing this interview because it’s probably been maybe north of a decade, probably longer at this point, that we’ve all known each other. So I’m really excited that you’re here and kind of sharing the Juno story. So why don’t you just give a quick introduction to each of yourselves, the story of Juno, and the specific problem that you’re both solving.

Founder Background & Why Juno Exists

Devon Tivona (02:31.246)

Sure. I can jump in. So Sam and I actually met in college and were working on separate startups at the time. And after co-working together post-graduating, we decided to build a company together. And it’s funny because this company that we’re building now owes itself to the previous company.

After lots of pivots, twists, and turns, we found ourselves in the corporate travel space with our last company, Pana, solving a very unique, interesting value proposition for enterprise customers, which was that for their employees, they have travel management tools and expense management tools, but for non-employees — people who the company is responsible for, whether those be new hires or speakers or partners or recruits or contractors, or in the healthcare space, healthcare professionals traveling on behalf of organizations — there are not good tools to solve for that pain.

And we found overwhelming demand for solving the gaps that exist in existing travel and expense tools. We built Pana to be basically the industry leader in the space. The pandemic hit, and we made the decision to sell the business, but all of our customers came back to us and said this is still a really valuable pain point post-pandemic when they turned back on their travel programs and started spending again. So Sam and I went back at it again with Juno. So Juno is a modern travel and expense platform built specifically for guests and non-employee travel workplaces.

Why Start Again After Selling to Coupa?

Dave Ambrose (04:10.303)

What made you decide to start a corporate travel business after you sell a company, which you kind of glossed over in terms of the story, but you sell a company to Coupa and you’re inside of a procurement business that at one point was public? Why do it again? Why? That’s the question I’ve always wondered.

Founder–Market Fit & “Unfinished Business”

Devon Tivona (04:35.886)

I think there are two reasons. One is very personal. Both Sam and I feel like we have unfinished business in this space. There was a core problem that we were just on the precipice of building up a very, very large company around that was durable, decades-long, that solved enduring problems in the procurement space. And then the pandemic happened.

So to be back at it again really feels like closure for Sam and me. But then secondarily, we had all this decade-and-change of procurement, travel and expense, and business spend management knowledge that is very unique and industry-specific, and we had nothing to do with it. And so this really feels like — a lot of times companies talk about product-market fit.

We feel like we have very unique founder-market fit in this space. So I guess you could say we are a very specifically shaped puzzle piece, and this is a very specific puzzle for us to go solve.

Go-To-Market: Direct Enterprise + Reseller Motion

Dave Ambrose (05:44.35)

Yeah, maybe take everyone through how the product kind of gets distributed. We won’t talk about the Pana days, but I think at the Juno days it’s a little bit different. But maybe walk through how you sell the product and maybe walk through a little bit kind of who’s the buyer that you’re selling to.

Sam Felsenthal (06:05.169)

Yeah, I can jump in here. So we have two different go-to-market motions at Juno, and one of them mirrors Pana. So we have a direct enterprise, very traditional sales motion, lots of hand-to-hand combat. We are selling to what’s called a travel manager. So this is somebody that typically is sitting in procurement or finance, and they own all things travel-related. So that’s negotiating their airline contracts, hotel contracts. If a company has an online booking tool, that’s kind of all within their purview. That role typically is in a company like a Fortune 500-sized company or quite large. And so we are doing a direct sales motion into that buyer. That’s the same thing we did at Pana.

On the Juno side though, we’ve also introduced a reseller motion. So in the travel world, there are these two travel management companies, or you can think of them as old-school travel agencies. There are two that own a lot of the market, and that’s BCD and AMEX GBT. Tons of acronyms in our space. And they’re like the true enterprise. There’s a tier below that with a long tail of these mid-market travel agencies. And it’s pretty hard to sell into their customer base.

They’ve got long relationships. Some of those customers are smaller. And so instead of us doing hand-to-hand combat there, we’ve established reseller partnerships with those mid-market TMCs. And so Juno is using both of those motions: the direct sales motion and actively scaling a reseller motion as well.

Vertical Pull & The Status Quo

Dave Ambrose (07:52.167)

And then for the majority of the market that you’re executing against, do you find that there are certain industries where there’s just a lot of buyer pull for Juno?

Sam Felsenthal (08:00.977)

Yeah, when we think about guest travel, there are a handful of horizontal use cases that apply to every single business. So that’ll be candidates, interns, new hires, contractors, meetings and events. And for that pool of travel, it’s typically 10 to 20% of any company’s travel spend. And so that’s cross-applicable to most businesses. The larger the business, the larger that pool of 10 to 20%.

But there are specific verticals where you go well beyond that 10 to 20%. And so that’s industries like education or the energy space — solar, wind, oil and gas, mining, construction — that in our industry is called ERM, energy, resource and marine. Sometimes that tends to change. And incentive travel. In those buckets it will be upwards of 50 to 80 percent of the company’s travel.

Dave Ambrose (09:02.865)

Wow. Wow. And then what’s the current solution in those cases where it’s like 50 to 80% of travel? Let’s guess. If they’re not using Juno today, what’s the other option?

Sam Felsenthal (09:15.121)

Yeah, it’s all offline manual workflows. So you’re picking up the phone, you’re calling, and you’re most likely talking to a travel agent. They’re doing the same things they were doing 30 years ago. It’s all the same. Literally, they’re using the same core infrastructure that they were using to book travel oftentimes 30 years ago, which is a terminal-like system called the GDS. So it’s all offline.

Devon Tivona (09:15.203)

It’s inside.

Ripping & Replacing vs Interface Layer Innovation

Dave Ambrose (09:46.94)

Do you think when you sell the product either through the direct motion or through the reseller motion — maybe on one end the reseller motion, the buyer there is the TMC, the travel management company that’s using the product and then selling it to others, and corporate customers direct could be in some of these GDS solutions that are out there.

Do you find that Juno is kind of ripping and replacing the GDS, or is it enhancing kind of the core workflow where the sale is actually not that difficult in terms of teaching a new workflow — it’s just more of enhancing it?

Because I think historically that’s always one of the challenges when you think of corporate travel and selling into the enterprise — kind of a lot of ripping and replacing, which is often very difficult from a start-up’s perspective in the sales motion.

Devon Tivona (10:35.146)

Yeah, I mean, I’m technical by background, and so I have some strong opinions about ripping and replacing versus continually iterating and improving. And I think the thing that most people don’t realize is that most innovation across most industries is incremental improvements, layers of abstraction on top of still some of the core gnarly old-school infrastructure that still exists.

If you look at the fintech space, ACH is still a CSV file that gets sent between banks every day to determine how money should be moved. But you use a mobile app and press a button and use Apple Pay in order to make that payment, built on top of all of those layers of abstraction.

So Juno’s not ripping and replacing some of the core underlying infrastructure such as the GDS, but we are making it much more accessible to build employee efficiency, customer efficiency and TMC efficiency on top of that GDS.

And actually, it’s better if you don’t rip and replace, because what people don’t realize is that technology that’s 30 years old also has 30 years of iterative improvements to wrap itself like a glove around the use cases and pain points that customers are experiencing. So we find you typically do more disruption by ripping and replacing, and instead we’re just incrementally enhancing.

Network Effects & Travel Infrastructure Complexity

Sam Felsenthal (13:13.359)

I think the other thing about travel — just to jump in really quickly — is there are both network effects that exist in traditional infrastructure in the travel space, as well as marketplace dynamics. And so you’re trying to disrupt two things that are traditionally very hard to disrupt.

On the content side, you think about the tens of thousands of hotels in the world. And in order for us to book stuff successfully, any travel program may want to book any of those hotels. So for us to rip and replace how hotels are booked globally, that becomes a pretty complex, hairy problem set to get all of those hotels in.

And then on the network effect side and kind of usage adoption, you have all these travel agents who have spent decades being trained on how to service travel very well in a very high-touch manner.

And ultimately in travel today, when something’s going wrong and you need help figuring something out, you want to be able to get on the phone and talk to somebody or have somebody help you solve those problems. AI over time might be an interesting disruptive force there, but we’re not fully there yet. And so you have these people that need to be armed with tooling to be able to help you do that well across air, hotel, car and all these things.

And that kind of infrastructure layer exists to enable them to do that.

Pre-AI vs Post-AI Building

Dave Ambrose (14:41.261)

So two more broader sets of questions before we wrap up. Before Juno, you built Pana, and this was before AI. And I think you built Pana in this world where we just started this concept of vertical software kind of becoming the system of record in a certain workflow for an employee. Lots of very specific integrations in a deep industry that you kind of get a lot of entrenchment, a lot of usage.

And now we live in a world where vertical software still exists, but on the interface layer, the interface layer has actually changed quite a bit, largely because of AI, ChatGPT, Claude and things like that. And now Juno is built in an AI era.

So I think it’s a really interesting conversation to have as two CEOs and founders of a business that was vertical software, that raised venture, that sold to a very large public company. And now you’re back at it again in an AI-native way.

What’s different about that from building a company pre-AI versus post-AI? And then on top of that, one of the things that I find so interesting about AI-native companies is that the rate of experimentation for employees is so much higher than that of pre-AI businesses. So maybe walk the audience through how you think about experimentation now in the Juno world.

AI as a Coordination Layer & Outcome Pricing

Devon Tivona (16:05.698)

Yeah, I can talk about the product, and maybe Sam, you can talk about how we think about experimentation. And apparently this is my day to be contrarian to your questions, but I’m actually going to start with what’s the same, because I think it sets the stage for what’s different.

Pre-AI and post-AI, what stays the same are the jobs to be done and the problem sets that we’re solving for customers. We view ourselves as a coordination platform for travel and for any travel journey, whether it be a guest journey or a meeting or event or your top salespeople going to an incentive trip.

There are 70, 80, 90 different decisions that need to be made along a trip. What airport should I fly into? What hotel should I book? How should I expense this? How should I pay for this?

In the pre-AI world, a lot of those pain points and problems our software could solve — lines of code, if-then statements — but some of them still required our customers to do work. You would use our product and point-and-click interfaces to make some of those decisions.

Now we just have another tool in the toolkit, which is AI, that can be trained to make relatively predictable and intelligent decisions about some of those pieces of the workflow.

So as we’re able to more deeply solve these pain points of travel coordination, the number of minutes or hours or days per trip that someone is spending thinking about logistics — when they’re not thinking about doing their job as a recruiter or as a salesperson or as a contractor — goes down. And we’re able to deliver higher efficiency and higher satisfaction.

And what’s really exciting about that is we’re able to actually tie that to our pricing model now.

In a pre-AI world, we said we charge you X dollars per trip. And if that trip turns into a disaster or if that trip goes well, we’re still going to charge you X dollars per trip.

Now we are so confident in delivering the outcome that we’re promising — which is that you don’t have to spend a lot of time thinking about your trip and your traveler is going to have a high level of satisfaction — that we survey our travelers at the end of the trip.

And if they score three stars or lower out of five stars, we don’t charge you for that trip because we did not accomplish our job — the job to be done, the reason why you’re hiring our software.

And that’s just — I think that’s really exciting because it means procurement folks can actually buy outcomes rather than just buying.

Experimentation in the AI Era

Dave Ambrose (18:50.88)

Yeah. Sam, what about experimentation for you?

Sam Felsenthal (18:55.163)

Yeah, I think one of the things that’s exciting about all this stuff in the AI world is the rate with which you can build across basically every job function.

So in the product and coding world, I think that’s gotten a lot of publicity, and everybody talks about Cursor and all these coding tools. And I think the rate at which you can ship something — and with different levels of stability — has increased significantly.

When people talk about vibe coding, we don’t believe in vibe coding for a product that needs to be super stable and scaled and all the rest. But it does allow you to start to get a feel for what a product could be.

So I think there are different levels to experimentation. Typically we’re thinking about whether this is a near-term experiment or a long-term thing that we need to be building toward.

On the product end side, there’s that piece of it. From a go-to-market perspective, basically in every vertical there’s now AI-specific tooling that you can use to run experiments. And we’re dabbling with all of it across everything.

Because ultimately we want to be able to build a highly efficient team who can do 10x what our old team could do in the Pana days.

Small Teams & Radical Transparency

Dave Ambrose (20:18.73)

Yeah. How do you share the learnings from the experiments inside the company? How do you share the learnings and the insights that you’ve gleaned from the experiments inside the company?

Sam Felsenthal (20:32.001)

So we’ve intentionally tried to keep our team small, and that’ll continue to be the case.

And I think that’ll actually be a pattern that we start to see across a lot of businesses — the headcount of companies may not grow as quickly, but the number of companies might be exponentially higher than what we saw before. And so that makes cross-pollination of successes and ideas and all the rest really easy.

We can talk to our team all the time, and everybody’s in Slack, and everybody can see what is happening.

So at Juno, we take the stance of being extremely transparent with the company — with everybody at the company — on everything we do. Engineers have full visibility into the sales funnel and the sales experiments that are being run. And our sales team has full visibility into the engineering experiments that are running, product experiments that are running and feedback that’s happening around that.

So rather than try to get cute or creative about how to share all this stuff, we’ve just taken the stance of let’s just share everything. And we’re hiring a small but high-agency team to go out and get the information they need from each other.

Closing & Contact Information

Dave Ambrose (21:51.006)

I agree with you that there are going to be more companies — an exponential number of companies — at a much smaller headcount base, which is structurally so different than what we saw historically in consumer internet and software, especially on the private side.

All right, guys, this is awesome.

Final question for you: for folks that want to get in touch and learn more about Juno, what’s the best contact information?

Devon Tivona (22:19.67)

Best contact information is to choose one of either Sam or myself and email. I’m devon@juno.travel, and Sam is sam@juno.travel.

Dave Ambrose (22:31.935)

Thanks guys. Thanks for joining today.