The sale looked complete. A customer had clicked, paid and received the thing. Then the transaction reversed direction. For an online merchant, a chargeback is commerce with a trapdoor: the revenue disappears, the product may already be gone, a fee arrives for company, and somebody must assemble a small legal brief before a bank's clock runs out. Avia Chen learned that rhythm while building ecommerce businesses with his brother, Ariel. They had software for almost everything. The exception was the task making them mutter at their screens.
At Babe Cosmetics, their cruelty-free beauty brand, the brothers knew the optimistic side of direct-to-consumer retail. The subscription operation grew quickly. Ariel later wrote that BabeBox reached 70,000 subscribers in its first year. The shop could add apps for upsells, subscriptions, marketing and payments. The ecosystem was a bright bazaar of useful little machines.
Chargebacks sat outside the bazaar. What began as a few disputes became hundreds, Ariel recalled, with first-party misuse cutting into more than 15 percent of bottom-line profit. Each case asked for a different bundle of proof: delivery confirmation, order records, customer messages, account activity. The rules varied by network and reason code. The work was both fussy and consequential, which is an excellent recipe for procrastination and a terrible one for cash flow.
“We couldn't find an existing one.”Avia and Ariel Chen, on the tool they went looking for
They sold the brand and kept the problem
Chen did not arrive at financial technology through a grand theory about banking. He arrived as an irritated customer. The brothers had been building things together since childhood; Chargeflow's own history puts their first joint venture when Avia was 12. By the time they were operating Babe, their partnership had acquired the usual family-business advantages and hazards: shared memory, compressed arguments and no need to explain where the bodies of old experiments were buried.
In 2021, they sold Babe Cosmetics and started Chargeflow. The first version lived where their own problem had lived, as a Shopify app. Its job was to pull together the transaction trail, understand the dispute category, generate a response and submit it. Instead of asking a merchant to become a part-time payments paralegal, the software would conduct the tedious scavenger hunt.
What used to occupy the merchant
This was a good problem for automation because its annoyance was structured. Card networks already supplied categories. Merchants already held evidence in their commerce, payment and customer systems. The missing piece was orchestration: fetching the right facts, arranging them persuasively and meeting the deadline. Chargeflow used machine learning and templates, with human review in its early described process, to prepare custom evidence. It also made an unusually legible pricing promise by charging on successful recoveries.
The product and the pitch were nearly the same sentence: connect your systems and stop spending attention on disputes. That fit Chen's blended role. He has been called chief marketing officer, product and marketing leader, and chief product officer in partner materials. The titles differ, but they point to the same territory. He works where the mechanism meets the explanation.
The unglamorous category has a good story
A chargeback is not naturally charming dinner conversation. Its vocabulary sounds as though it was designed by a committee that feared adjectives: representment, reason codes, retrievals, pre-dispute alerts. Chen's marketing challenge was to preserve the complexity while making the stakes plain. His solution has been to speak as the former merchant in the room.
He writes about checkout friction, privacy changes, customer experience and payment disputes with the practical impatience of someone who has had to keep the store moving. His “Ask Avia” format turns that operating history into advice. The useful tension in his public work is between growth and its consequences. More transactions are welcome. More transactions also create more opportunities for confusion, fraud and customers who decide the bank's dispute button is easier than the merchant's support desk.
That perspective also made the story expandable. Chargeflow moved beyond responding after the fact. Alerts aimed to intercept disputes before they hardened into chargebacks. Insights gave merchants a view of their risk. Connect let payment providers and other platforms embed the system. Prevent moved earlier still, toward detecting friendly fraud before the post-purchase mess. The company was turning one chore into a map of the whole problem.
Capital followed the widening brief. In August 2023, Chargeflow announced an $11 million round led by OpenView, on top of a previously undisclosed $3 million pre-seed. At the time it had roughly 40 employees and plans to double. In November 2025, the company announced $25 million in equity led by Viola Growth and $10 million in debt. Reported total funding reached $49 million.
The business had changed scale as well as shape. Chargeflow said revenue had more than tripled over the preceding year, its customer base had more than doubled, and more than 15,000 businesses used the platform. It named companies including Fanatics, WHOOP and Sweetgreen. The staff count reported at the fundraise was 160, spread across Israel, the United States, the Philippines and Poland. The lonely Shopify app had become a distributed company explaining the subtleties of disputed card payments across time zones.
Babe Cosmetics grows, and chargebacks expose a costly gap in the ecommerce tool kit.
The brothers sell Babe and launch Chargeflow.
An $11 million seed round takes disclosed funding to $14 million.
Insights, Connect and Prevent extend the company from recovery toward visibility, embedding and prevention.
A $35 million financing package takes reported total capital to $49 million.
A sibling company with a long memory
Founding with a brother makes for easy folklore. The more useful detail is accumulated context. Avia and Ariel had watched the same customer behavior, paid the same penalties and stared at the same holes in their reporting. They did not need a research project to prove that the task was miserable. They needed to decide whether their misery was sufficiently common.
Their roles create a revealing division. Ariel, the CEO, is usually the public voice on company operations and the mechanics of the market. Avia translates the lived problem across product, brand and education. In a comment after the 2025 financing, Ariel described his brother as visionary, innovative, ambitious and someone who “lead[s] from the heart.” Praise between brothers should always be discounted slightly, like an invoice paid suspiciously fast. Yet the public evidence supports at least the connective part: Avia's work joins customer frustration, product design and category language.
A Stripe customer story offers a small example. Chen argued that becoming an official partner created trust and let customers stay inside the Stripe Dashboard instead of bouncing between systems. It is an unflashy observation with strategic weight. Software wins not only by doing a job well, but by appearing where the job already happens. Every avoided login is a tiny act of mercy.
“The job is not finished. It's only the beginning.”Chen after Chargeflow's 2025 financing
Making the dispute disappear
Chen's stated ambition is not to make merchants better at paperwork. It is to make the paperwork recede. In his 2025 financing announcement, he described a model in which disputes are prevented before they begin, evidence is compiled and submitted automatically, and operational blind spots surface on their own. The chargeback becomes background workflow rather than weekly crisis.
That ambition contains a paradox. Chargeflow must make the category important enough to buy, then make its daily presence forgettable. A successful dispute platform should feel less like a cockpit and more like plumbing. Nobody throws a party because the drain worked. They notice when it did not.
By 2026, Chen was also thinking about how expertise travels through AI systems. He argued publicly that years of publishing proprietary data, detailed articles and genuinely useful forum answers helped Chargeflow become frequently cited by large language models. Whatever the precise ranking, the strategy is consistent with his career. Take a murky system. Explain it in the language of the person stuck inside it. Leave a trail clear enough for both people and machines to follow.
The enduring part of Chen's story is smaller than a financing round and more portable than fintech. He and his brother did not merely spot an inconvenience. They had paid for it, in fees, hours and damaged margins. Selling Babe Cosmetics closed one company, but its ugliest operational lesson came with them. They turned that lesson into a product, then a platform, then an attempt to remove the problem before a merchant ever sees it.
A refund is supposed to return money. In Chen's case, the troublesome ones returned something else: a company idea. Commerce is generous that way. It sends its sharpest lessons after checkout.