Catherine Jue had spent years working in the technical field, studying things like natural language processing and artificial intelligence. But when handling the accounting for her father’s small business, Jue discovered a problem: For the most part, everything was still being done manually. She realized there needed to be a way to automate bookkeeping so that companies are able to focus more on products and customers, improving the accuracy of their finances without breaking the bank.

This has given rise to Sway Finance, a fintech startup aimed at automating accounting for Software as a Service (SaaS) businesses. Together with cofounder Ashwin Kumar, Jue aims to curate all the financial data from bank accounts, Stripe, Gusto, Expensify, and other services to produce real-time reports that let businesses understand how well they’re performing.

Primarily accessible through Slack, this Y Combinator-backed company has evolved its bot from one that simply provides balances and transactional history to one that does more full-fledged accounting. With its integrations into third-party platforms, Sway Finance automatically generates income statements and balance sheets in real time, no longer requiring accountants to wait until the end of the month. It’ll also take care of other financial matters, such as bank reconciliation and journal entry adjustments, so that companies are up-to-date and can easily prepare for tax filings each quarter.

sway_categorize

We’re told that you can query data by date range with a command line request, such as “What were the previous day’s sales?” There is no web dashboard right now, but Sway Finance is working on it.

The company started out offering simple summaries before evolving into something with some “creative pizazz.”

The product is specifically geared toward SaaS businesses, something Kumar shared is because accounting for these companies is relatively straightforward. In its current state, Sway Finance may not be the right solution for everyone, especially companies with more complex financials.

Kumar and Jue believe their competitors are people who are trying to do accounting themselves or are paying traditional firms for these services. “There’s nothing in SaaS right now with machine learning,” Kumar said.

It’s this technology that is Sway Finance’s secret sauce. Machine learning “works best in labeled data sets where you have tons of data which humans have labeled,” Kumar explained. In accounting, “you categorize transactions, a tedious thing, and can look at thousands or millions of transactions, learning how to categorize them and do it better.” He claimed that those who use software like Mint.com often complain about transactions being categorized incorrectly, but Sway Finance has found a way to get a 90 percent accuracy rating. Much of this is because of the evolving nature of its algorithm, which is constantly learning and training itself as more data gets entered into the system.

Sway Finance has 10 companies signed up for its pilot program and charges users $99 per month. There’s no maintenance required, as it’s a “set it and forget it” type of service. More integrations will be added in the future, and there’s the possibility of rolling out a developer API and an analytics feature. Kumar said that the company is exploring how to track other financial metrics, such as Monthly Recurring Revenue (MRR), churn rate, number of users, and profit per customer.