The cursor hovered over Send. On the other side of the screen were hundreds of finance executives who had no reason to answer Caitlin Leksana. She was not selling them anything, at least not yet. She was asking for a few minutes and an education. The distinction did little to calm her. “I was scared,” she later recalled. “Seriously, it's nerve-wracking.” Then she sent the messages anyway.
Leksana and her co-founder, Timmy Galvin, had reached the kind of moment that makes a clean founder biography impossible. Their first startup, Carma, sold community and marketing software to crypto companies. They built a product, won enterprise customers and watched the market underneath those customers fall apart. Churn arrived. Prospects wanted custom work. The pair kept going, then finally had the conversation they had been avoiding: this was not the company they wanted to keep building.
But Carma had left them a clue. Even with roughly ten customers, they were color-coding spreadsheets to track invoices and payments. The work jumped among contracts, email, accounting software and bank records. A company could make a sale and still struggle with the seemingly ordinary business of getting the cash into its account. The coping mechanism looked small. The problem behind it did not.
“Every cold email someone responded to energized me even more, because it meant we were finding a problem without a solution.”Caitlin Leksana
The education of an outsider
Neither founder was an accountant. Leksana framed their options with dry efficiency: they could become accountants themselves, or they could talk to people who already understood the work. The second route sounded faster. Her cold messages asked to learn rather than pitch. Dozens of CFOs replied. Calls booked for 20 minutes opened into tours of aging reports, remittance details, customer portals and the rituals teams used to discover why an invoice remained unpaid.
The conversations supplied more than validation. They supplied language. On one call, Leksana asked a CFO how she handled a certain process. The CFO looked puzzled and asked whether she meant “dunning,” the sequence of reminders sent to collect overdue payments. Leksana did not know the word. She was embarrassed. Later she laughed about the exchange because dunning became a core feature of Fazeshift.
Some founders treat unfamiliarity as a liability to conceal. Leksana turned it into a research method. Her outsider status forced the basic questions that insiders might skip. Why did one payment require visits to an ERP, a CRM, a bank portal and an inbox? Why did a customer need a particular PDF attached in a proprietary system? Why had software automated the neat center of the process while leaving people to carry all the odd cases between systems?
A career that compounds
Leksana arrived at that messy software problem through mechanical engineering. At Georgia Tech she studied control systems, robotics, nanoengineering and electrochemistry, earning bachelor's and master's degrees. She also performed with the color guard, eventually serving as captain, taught a mechanical engineering course and participated in the Consulting Club. The mix is revealing without needing to be tidy: technical systems, public performance and an interest in how organizations make decisions.
After graduating, she joined Boston Consulting Group in Atlanta. For two years she advised technology companies on product, operations, pricing and go-to-market strategy. Then came Harvard Business School, where she met Galvin as a section-mate. He brought computer science training from MIT and experience as a nuclear submarine officer. They graduated in 2022 and started Carma together.
From a distance, the turns look abrupt: machines, consulting, crypto communities, receivables. Up close, the skills accumulate. Engineering offered a way to work through unbounded systems. Consulting added a vocabulary for product and pricing. Carma delivered the operational bruise. Harvard introduced the partnership. “Mechanical engineers are good at not getting discouraged,” Leksana has said. “They can think through ambiguity and complexity and figure out how to focus on what's most important.”
The snowflake in the back office
Accounts payable can be standardized by the company paying the bill. Accounts receivable is shaped by everyone else. Every customer has its own rules, portals, approval chains and ideas about what belongs on an invoice. Leksana calls it a “snowflake” problem. The phrase makes an old back-office function sound almost charming. In practice it means fragmented data, patient follow-up and countless opportunities for a payment to lose its context.
Fazeshift sits across the tools a finance team already uses. Its agents are designed to execute work spanning invoicing, collections, payment matching and reconciliation. The company describes the product as a context or control layer: something that can see the contract in one system, the conversation in another and the eventual deposit somewhere else. The aim is to automate a chain of work, not merely make one link faster.
Fazeshift entered Y Combinator's Summer 2024 batch. In January 2025 it announced a $4 million seed round led by Gradient Ventures. By May 2026 the company said revenue had grown twelvefold in a year, with dozens of enterprise customers including eight unicorns and its first public company. That month it announced a $17 million Series A led by F-Prime Capital, bringing reported total funding to $22 million.
The milestones arrived in public as a steady sequence rather than a single reveal. Fazeshift took first place at the Finance & Accounting Tech Expo pitch competition in 2024. The following spring it was selected as a finalist for the SXSW Pitch competition in the category that bundled enterprise software, smart data, fintech and the future of work. These stages gave Leksana a compact version of the job she had already been doing on cold calls: explain why an ignored process deserves attention, then make the complexity legible without sanding it away.
Her comments after a female-founder panel offer a useful view of how she measures progress. Money, she wrote, is “oxygen, not success.” Success comes from a strong team and a product people love and use regularly. She also repeated a two-pizza analogy about small teams: when capital is scarce, founders cannot casually hire around every problem. They have to stay close to the work, resist bureaucracy and decide which problem matters enough to solve. The remark is especially pointed for a company selling automation. Fazeshift's pitch is not that finance work lacks value. It is that skilled people should not spend their time acting as the connective tissue between software systems.
That distinction keeps the product story grounded. The company is not trying to make the office of the CFO look futuristic for its own sake. It is trying to remove the recurring search for context: which invoice a deposit belongs to, whether a customer received the right attachment, what promise was made in an email and who should act when the ordinary workflow breaks. The work sounds procedural until cash is late. Then it becomes a company-wide concern.
The numbers matter, but they do not erase the deliberately unglamorous starting point. A CFO told Leksana she spent three hours every Friday handling invoices. Another reaction was even more direct: please build something that solves this. Several of the people who explained the problem during those early calls became first customers when Fazeshift launched. Customer discovery did not sit before the company. It folded into the company.
“What's the worst that can happen? Someone doesn't respond. What's the second worst? You talk to people and disprove your startup idea. Either way, you've made progress.”Caitlin Leksana
From doing to reviewing
Leksana's larger argument is about the shape of finance work. Teams today bridge systems by hand. As agents take on those transitions, people move from doing every step to reviewing outcomes, managing exceptions and setting governance. That shift requires trust because an agent touching invoices and cash cannot operate like a clever autocomplete. Finance teams need auditability, accuracy and a clear way to stay in control.
Her ambition now extends beyond receivables. She has described a broader CFO suite and a future of autonomous finance in which AI executes core operational work while human teams focus on agent management, strategic decisions and governance. It is a large destination reached through unusually specific terrain: missing remittance advice, unusual billing terms, attachments, portals and the customer who insists that Part A and Part B arrive as separate PDFs.
In 2026, Leksana was named to the 100 Women in AI list. The recognition arrives after a period of visible company growth, but her most useful story remains the one before the metrics. A founder admitted what she did not know. She sent messages while nervous. She let finance leaders replace her vocabulary with theirs. She and her co-founder stopped protecting a company that no longer worked and studied the annoying task it left behind.
There is a temptation to make a pivot sound like a flash of insight. This one looks more like accumulated evidence. Ten customers made a spreadsheet buckle. A market collapse forced a hard conversation. A hundred strangers described the same fracture from different angles. The startup appeared gradually, in replies, corrections and Friday afternoons spent chasing invoices. Leksana did not set out to make accounts receivable interesting. She paid enough attention to notice that it already was.