#2288 She had people do AI’s work
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
Helen Hastings, founder and CEO of Quanta, built her AI-powered accounting platform by first having humans perform the work that would eventually be automated. In this podcast, she reveals how she validated her idea to replace QuickBooks by shadowing bookkeepers and understanding their pain points firsthand. Hastings leveraged her background building financial ledgers at companies like Affirm to create a modern accounting system that provides real-time financial data instead of the traditional month-long delays.
- Key themes include using human workers to prototype AI solutions, the critical importance of clean data architecture, and the strategic decision to focus only on software companies that could be fully automated rather than going horizontal across all business types.
Speakers
Helen Hastings
Helen Hastings is the founder and CEO of Quanta, an AI-powered accounting platform built for modern software and services companies. Before Quanta, she was a software engineer at Affirm for almost six years, specializing in building financial ledgers and systems of record, joining when the company had around 100 people and staying through their IPO and growth to over 2,000 employees. She's also worked at Google and NerdWallet, bringing deep fintech and infrastructure experience to solving accounting's hardest problems.
Key Takeaways
Use Humans to Prototype AI Solutions
Helen's most innovative strategy was having humans perform the work that AI would eventually automate. (00:25) This approach helped her understand exactly what needed to be automated and provided a bridge for customers who weren't ready to trust fully automated systems. By having humans read and categorize expenses, analyze vendor relationships, and process financial data, she could validate that the work could be systematized before building the technology. This "human in the loop" approach is particularly valuable in AI-era startups where you need to prove the concept works before investing heavily in automation.
Focus on Clean Data Architecture First
The foundation of Quanta's success lies in building clean, structured data storage before applying AI on top. (36:02) Helen emphasizes that "if you're taking ChatGPT and putting on messy data, you can only get garbage out if you put garbage in." Traditional accounting systems like QuickBooks allow users to make changes that break data integrity, but Quanta built redundant checks and continuous reconciliation from the ground up. This creates a reliable foundation that enables advanced AI analysis and real-time reporting.
Start Narrow, Then Expand Strategically
Rather than trying to serve all businesses like QuickBooks, Helen deliberately chose to only work with early-stage software companies that fit within their automation capabilities. (27:04) This meant saying no to many potential customers, including nonprofits and businesses with physical inventory. By focusing on companies that use modern financial tools like Stripe, Mercury, and Rippling, Quanta could achieve full automation and deliver exceptional service quality before expanding to other business models.
Validate Willingness to Pay Early and Often
Helen conducted extensive user research, reaching out cold on LinkedIn to finance managers and accounting professionals to understand their pain points and validate demand. (16:25) She emphasizes asking people if they're willing to pay for a solution, noting "I think it's essential to start a business." This validation helped her understand not just that the product would solve pain, but how customers would perceive it and what would drive them to actually purchase and switch from existing solutions.
Build Real-Time Financial Visibility
Traditional bookkeeping creates dangerous delays where businesses don't understand their finances until weeks or months later. (14:00) Helen discovered that companies often don't have their financial data ready until well into the following month, sometimes taking until the end of the next month to start processing. Quanta's real-time approach allows companies to make decisions on a day-to-day basis, which is particularly critical in the AI era where companies need to track volatile costs and usage-based pricing models.
Statistics & Facts
- Helen joined Affirm when it had around 100 people and stayed through its growth to over 2,000 people and IPO. (11:23) This experience building financial ledgers at scale gave her the technical foundation to understand what was needed to replace traditional accounting systems.
- Quanta raised $20 million in total funding, with a recent $15 million Series A led by Excel. (36:45) The company currently serves just under 100 customers, demonstrating solid traction in their focused market segment.
- Companies using traditional bookkeeping services often wait until well into the following month to receive their financial data, with some not getting October data until December when recording in early December. (14:40) This delay makes the financial information largely useless for decision-making in fast-moving startups.
