Workato × Tiro × Bloom: How a Korean Became No. 1 in Global Sales

We ran a Bloom event with Yejin from Workato and Yeoul from ThePlato, the company behind Tiro.

At our events people often ask, "How do you get your speakers?" The truth is there's a story behind each one.

With the Tiro team, it started when we began running English sessions from our second event. We suddenly needed an AI interpretation service, and Tiro was the first to reach out and help. Because of that, we now summarize almost every event with Tiro, and we use its AI interpretation every time an overseas company joins us for a session. They've always given us a stable service and careful support, and I'm truly grateful.

With Workato, I'd wanted to do an event together in Korea for a long time, but their global schedule is packed and the timing was hard to line up. Then, on a business trip to Singapore, we planned a company event there, and Workato very kindly opened up their office as a venue for us. That gave us the chance to have deep conversations with the Korean employees in Singapore and hear firsthand about Workato's enterprise AI solutions and their overseas sales stories.

Thank you again to both speakers for sharing their experience and know-how so honestly and generously. And thank you to everyone who came.

What it means to turn conversations into assets

The opening was Sangcheol Kim, CTO of ThePlato, which makes Tiro. He goes by Yeoul, and I asked him to explain what Tiro is in his own words rather than a sales pitch.

It started as an AI meeting notes tool, and now it's positioned as a product that turns conversations into assets. Up until about a year ago, plenty of people were willing to pay just for accurate meeting notes. Now they're at the stage of asking what on earth they're supposed to do with all the notes they've piled up. If something is an asset, then once you've built it up, you should be able to buy a house or a car with it. The conversation data we accumulate should work the same way, he said. It began when GPT-4 was taking off. He had a hunch that voice would come after LLMs, thought it would be great to translate business English in real time, and built a real-time demo that same day. That was the start of Tiro. Right now B2C makes up 60 to 70 percent of revenue, but B2B has been growing fast lately.

The threshold everyone else avoids is the moat

I asked if he remembered their first B2B customer. He said he never will. It was April of last year, when every large Korean company was setting up an AX (AI transformation) division. A team lead who'd been using Tiro personally as a B2C customer reached out first when talk came up of trying AX at the organization level. They were lucky that meeting notes is a product where results are very easy to measure. If specialists spend this many hours, that's this much in labor cost, and adopting meeting notes saves this much. The math comes out immediately.

The obstacle was security. Meeting notes data is extremely sensitive, so it's an area startups rarely dare to break into. Yeoul's explanation was crisp: B2C runs on "try it, and pay if you like it," while B2B runs on filtering out reasons to say no. No security certification, cut. This doesn't work either, cut. That happens regardless of how satisfied anyone is with the product. They didn't have a security certification at the time, so the customer handed them a security review checklist with about 200 items. They passed every one with documented evidence and promised to get certified within six months. He then prepared for ISO 27001 and SOC 2 Type 2 at the same time, by himself.

That's a big challenge for a startup, so why did they decide to go for it? Once he started coding with GPT-4, he realized early on that the skill of building software fast would stop being a moat. So where was their moat? Maybe it was exactly this security review they were struggling through with the checklist, and the whole process of breaking into large companies. He figured that once you crack one, the rest come pouring in. And sure enough, once "no security certification, cut" was solved, a compounding effect kicked in: "These guys use it? Okay." I thought this was exactly right. Anyone can build a meeting notes app now, but not many companies take on the hurdles of selling to enterprises head-on.

Why they moved to Japan

They're a Korean startup, but they raised money in Japan first. They had no prior ties to Japan either. The company has a goal of building an export business that helps solve Korea's low birth rate, so they had to earn foreign currency from day one, which meant the product had to be global from day one. Narrowing down to countries with a small time difference, quick travel and a similar culture left Japan and the Middle East. The reason they dropped the Middle East was funny: the thought of building for a language written right to left made them really not want to do it.

And this team doesn't visit Japan on business trips. They go live there. One co-founder moved to Tokyo this spring, and everyone laughed when he said he'd found out about the move before his parents did. Even the way they learn Japanese is very them. Instead of signing up for classes, they go to Japan and make plans with Japanese people. When I asked why go that far, he said it's the same logic they've used to grow the product. You don't build something and blast it out. You make your first few customers extremely happy so they spread it. And to make your first customers extremely happy, you have to live with them.

Once they got there, Japan turned out to be very different from what they'd assumed. It's extremely conservative and cautious. Large Korean companies adopt AI fast and even talk about MCP, but in Japan, even fairly big companies still do most of their work in Excel. The sources people trust for new technology are different too. Koreans like short, punchy ads; in Japan they need to be dense and full of explanation, and people don't really accept something unless it has Japanese references. Put the other way, once you're locked in, it lasts a long time. Since the time and cost of landing a first customer are high, they decided it was better to settle there than to keep flying back and forth. The US was a different story. Right now American customers use Tiro for its Asian language support. Long term, though, it's a country where if you go just a little south of LA, lots of people ask what a notetaker even is, and he said he wants to get into a market where nobody knows him but everyone uses Tiro. He also compared the AI notetaker market's maturity to people asking whether Netflix and EBS (Korea's public educational broadcaster) are the same kind of internet company. That one stuck.

Splitting agents by person, not by domain

I also asked how the company works in an AI-native way. Usually people split agents by domain, like customer support or marketing. This team copied people. A company runs on people who are given roles, so each person copies themselves, delegates work like raising a Pokémon, and steps out. To do that, you have to define what you do clearly in words and break it into pieces that can be assembled. Solving that problem, which is really a question of philosophy of language, is the core of an AI-native organization, he said.

The agent structure was concrete too. The wiki, or memory layer, holds what's unique to you, like your speaking style and preferences. Every meeting at the company is recorded with Tiro, and the agents pull it all in once a day. On top of that sits an ontology layer that holds the rules attached to a role, like CTO or marketing. Since someone else might take over a role, rules that don't belong to an individual are defined separately, and the two layers get assembled. When asked where to start, he pointed out that Claude Code started delivering value because of a strong harness and a lot of context fed into it. So for a business, start by recording every conversation where the organization's judgment shows up, and regularly feed it to agents in a form they can digest. I record almost everything myself whenever I can. Having the information and not having it are different things.

His answers to audience questions were just as plain. They don't build their own models. They use APIs for everything and route to whichever model performs best at the time, and they lean heavily on ElevenLabs for voice. The most important metric is activation. Signing up doesn't mean someone's using it, so they define an active user as someone who records 10 minutes or more after signing up. They treat token costs for free users as a marketing expense and give them the first 300 minutes.

She joined because she liked flying

The main session was Yejin Son, head of Workato Korea. That day she'd gone to bed at 5:30 a.m. and flown in on a 10:30 a.m. flight. Her analogy for what Workato does was good. In a Marvel movie, when Iron Man says "get me the info from here, get me the info from there," a very smart AI pulls from multiple systems. Workato is the solution that handles that pulling in an enterprise context. Connecting one API with Claude Code is one thing, but an enterprise has 100 or 200 systems.

Her career started almost by accident. Born and raised in Korea, she finished college there and happened to land a Boeing internship, which was her first time in the US. Six months later she asked to be converted to full-time, but that's exactly when Trump was first elected and the visa route closed. With a recommendation letter from leadership and no other information, she moved to Boeing in Singapore and worked there for nearly six years. When I asked why she joined Boeing, she said simply that she loved to travel, so she liked flying, and it really was a company where she got to fly a lot. She did technical sales, selling software to every airline in Asia Pacific: systems for captains, systems for flight crew, fuel optimization tools and so on. I asked whether being Boeing meant always having the upper hand. She admitted frankly that those were her arrogant years. She was clearly going there to sell, but for a long time she went in with the attitude of going to teach.

About six years in, she realized it was a really good workplace where you had plenty of chances to get promoted even if you just did an okay job, and at the same time she felt like she was falling behind. She decided to change jobs at the start of COVID in 2020, and what she did next was fun. She'd vaguely heard there were things called Series A, B and C, so she pulled a list of Series C, D and E companies in Singapore and went through it. That's how she ended up at Workato. Contrary to what you'd expect, that it'd be tough, it fit her perfectly. Nobody told her what to do; they just told her to bring results. That was the exact opposite of her previous ten years, where results were set in advance and you followed the process, and that's why it fit.

Listening to global sales calls on the treadmill

At first she was a solutions consultant covering all of APAC. The next year, Korean sales started with: "You speak Korean, so you could probably sell. Just try it." She's the type who digs relentlessly into what she doesn't know, so whenever she gets a new task, she starts by pinning down what she's missing. So the first thing she did was listen to recordings of sales calls from other countries stored in the company's meeting tool. She listened while running on the treadmill in the morning, while eating, in taxis. It was COVID and every call was recorded, which actually made it easier, she said. When I learned sales, following a senior colleague around and watching from the side made a big difference. This was the first time I'd heard of someone soaking that up through recordings.

Selling in Korea meant a business trip every month, and what she did there was look for the problem only Workato could solve. A customer might have about 500 problems, and 490 of them can be solved without Workato. Of the remaining 10, maybe one is worth paying Workato's price for, and finding that one is hard. The approach was different too. Walk in saying "Hello, this is Workato" and people at big companies won't even look at you. Instead she sought out the hidden experts, people with long careers who wanted to build something with this kind of solution. Those people were surprisingly well connected in the startup scene, and talking with them always led to an introduction or two.

After she had a baby, her schedule got even tighter. She'd put her child to sleep, catch a night flight, close a deal and come back.

"Why do you think you can't do a one-million-dollar deal?"

Then I asked about becoming No. 1 in global sales. When she hit No. 1, the Singapore office had five people covering Korea, and they set out to change how those five thought. SaaS companies usually call a $100,000 to $200,000 deal very big. Her leader at the time poked at her pride and asked:

"But why do you think you can't do a one-million-dollar deal?"

That's about 1.5 billion won. But it's still one signature on a contract. Taking that challenge, the team started working backward. For a customer to spend 1.5 billion won, it needs to deliver 15 to 30 billion won or more in impact to be worth it. Where do you find that problem? The answer was in their conversations with existing customers. A customer who'd adopted Workato for a small use case started asking about all sorts of things once trust built up. "Yejin, have you heard of this vendor?" Instead of saying "Oh, I know someone good" and leaving it there, she asked, "Why are you asking?" and set up a meeting. That was a huge turning point. After talking it through, the customer concluded on their own that Workato was the solution to their real problem. She added an observation: Western sales is based on "solve my problem," but in Korea and Asia, defining the problem can come down to how well the salesperson understands and supports the personal agenda of someone who wants to move up at that company.

When I asked what the company does for you after a deal that big, her answer surprised me. Big rewards come with big responsibility, and it was really hard. A customer investing that much means expectations are just as high, and since it was a scale their sales org had never done before, there was a lot of responsibility to carry for delivery. Selling isn't the end; it's where things start. She also talked about what she learned when they fell short of expectations. At first she thought spending more time with the customer herself was how she could help. Then she realized that quickly finding the people who can solve the problem and putting them in the right places is also leadership.

The decision to set up a Korea office took four years. It was a process of being asked to prove things to the point where it felt petty, and to make the case that an office was needed, they absolutely had to have a big deal, so she was obsessed with that. The one-million-dollar deal wasn't just a number. It was the justification for opening the office.

The real bottleneck in Korean AX

In the first half of this year alone she had first meetings with 50 enterprise customers, and the AX struggles she saw at large Korean companies came down to defining what they want to do and change management. Sales AX, supply chain AX, turning data into assets, all fine, but defining the problem, what to do to get what effect, is hard. That's one. The one that got more nods in the room was change management. It's not just Japan that runs on Excel. Mid-size Korean companies all work in Excel too.

"The AX we picture is building something with Claude Code. The AX they picture is having to think a little less in the Excel window they stare at every day."

If you don't understand that spectrum and just show up with your solution saying "you can do this in a flash with Claude Code," you'll have plenty of meetings and never get a contract. For them it's so far off that the reaction is, it's like watching Star Wars.

Her answer about work-life balance was the most honest. "Honestly, I'm curious too. I have no idea how everyone else survives." She said she'd burned out and is someone who keeps running without having gotten over it, so she didn't have good advice. But she added that having coworkers who help you bounce back from burnout, motivation and small goals lays the groundwork for getting over it, very slowly. It sounded like someone who knows it doesn't just go away when you switch jobs.

Last, I asked how global companies see the Korean market right now. When the US looks at East Asia, there's a tendency to leave out China, which leaves Korea, Japan and Australia. SaaS companies that used to go hard on Japan are now looking at Korea's purchasing power in the numbers.

"These people don't move unless they see it in the numbers."

The most memorable rejection you've had recently

At the roundtables, we asked about the most memorable rejection people had gotten recently, and how they could do more with AI in sales and marketing.

I didn't expect the topic of rejection to land so well. Someone who teaches AI got complaints that it was too hard, made it easier, and later got requests to make it more technical. Their takeaway: if they'd clearly shown at the start what this technology can produce, the hard material would've been seen as worth learning. The boring intro meant as a build-up actually became the barrier. Everyday rejections poured out too: being ghosted, compliance notices getting rejected over and over, even a landlord getting turned down by a tenant. One table had no salespeople, so they broadened the idea. If every exchange of value and every act of persuasion is sales, then even a compliance officer is, in a broad sense, marketing themselves.

Someone driving internal AX said the developers push back because they feel their work is being encroached on, so they keep at it by using research to give those developers credit for their experience. Someone else talked about a marketing automation workflow where you just enter a URL and it runs everything from analysis to ad buying to reporting. The conclusion was good: even with everything automated, the important calls, like which strengths to lean on or whether to raise ad spend or cut it, stay with a person to the end. The funniest rejection was someone's sister-in-law telling them not to start anything new. That table's conclusion was firm: AI can write the algorithm, but it's people who move people. Win the sale with sincerity first, then use AI when it's time to scale up.

A failure story from overseas technical sales was specific. They pitched a Vietnamese conglomerate, but the other side wanted a penetration testing tool, while their strength was static analysis. The two sides wanted different things. They had meetings, dinners and drinks, and it still didn't happen. If they did it again, they said, the moment the fit was off they'd stop dragging it out and find another customer. Timing came up at two tables. A customer who'd held off three months earlier because a certain feature was missing heard it had launched, and came back saying they'd been unhappy with a competing solution anyway and would take another look. The conclusion: B2B is a matter of timing. There was also a warning that with so much AI-made content out there, people now see certain keywords and think "AI wrote this too," which can backfire where sincerity matters.

One table shared numbers: 400 cold emails, 16 replies, eight of those still in progress, and the rest all rejections. There was also an incident where a teammate's AI glasses caught a phishing meeting link from someone impersonating a former buyer, which saved them from real trouble. Someone visiting from New York described the pace there: business meetups happen on Luma, and investors watch a 30-second demo and go straight home. At the last table, a developer told a story about being turned down for a tool purchase because of budget. The salesperson leading that table answered: "Sometimes deals fall through for me because of budget too. Hearing this, I thought, oh, that's how the person who wanted to buy it must have felt, and it hurt for me too." Some things you only learn when the selling side and the buying side sit at the same table.

In the end, both of them won in the same place

The two speakers that night were headed in opposite directions. One is going from Korea out to Japan and the US. The other opened Korea for an American company. But they won in the same place. Tiro decided there's no moat in the skill of building software and made its moat out of the security reviews and enterprise sales process that others avoid. Workato spent its time finding the one problem out of 500 that only it could solve. Both won at the threshold others don't want to cross, rather than on product features. And both ended up going to live in the market. One team's co-founder moved to Tokyo; the other speaker put her child to sleep and caught night flights. The line about having to live with your first few customers to make them extremely happy ran through both sessions.

Originally written in Korean on Brunch, August 25, 2026.