🎧 Highlights:
[00:00:00] Humans doing the work of AI — before AI existed
[00:01:12] Why accounting is mostly about language, not numbers
[00:02:33] Shadowing bookkeepers to find automation opportunities
[00:06:00] Manual work Quanta knew software had to replace
[00:07:30] Why building on top of legacy systems wasn’t enough
[00:08:24] Rebuilding the ledger from the ground up
[00:10:12] Continuous reconciliation vs. monthly closes
[00:11:24] From Affirm to founding Quanta
[00:13:30] Why delayed financials are useless for startups
[00:16:03] Validating willingness to pay before building
[00:17:42] Using humans for the “last mile” while automating the rest
[00:20:15] Solving trust and data-ownership concerns
[00:22:48] Why most QuickBooks challengers failed
[00:26:33] Saying no to customers to protect quality
[00:33:36] Why AI makes real-time margins mandatory
[00:36:45] Raising $15M Series A ($20M total)
[00:37:21] Prism: asking your financials questions in plain English

In this episode, Andrew Warner interviews Helen Hastings, founder of Quanta, an AI-powered accounting platform built for modern software companies.

Before AI could reliably understand financial data, Helen and her team had humans doing what AI does today — reading receipts, interpreting memos, categorizing transactions, and reconciling books by hand. That hands-on approach helped her uncover where automation really mattered, leading to a ground-up rebuild of accounting software that works in near real time.

Helen shares how Quanta replaces legacy systems by owning the data end-to-end, combining clean ledgers, continuous reconciliation, and AI-powered analysis — and why this approach helped the company raise $15M in Series A funding (over $20M total) and land nearly 100 customers so far.