Building Foresight: What I Built, What I've Learned, and Where Things Stand
I've been building Foresight for the past several months, an AI-powered FP&A platform for QuickBooks SMBs. Here's an honest look at where things are.
What I built:
The full product covers everything a CFO would do: live financial statements, 12-week cash flow planning, 12-month three-statement forecasting, variance analysis, rolling forecast, and hiring and scenario planners. On top of that sits Foresight Intelligence, an AI layer that reads your actual QuickBooks data and tells you what it means and what to do about it. There's also Ask Foresight, a conversational AI CFO that lets you ask plain English questions about your business and get real answers based on your actual QuickBooks and Foresight data. And a firm tier so CPAs and bookkeepers can run it across their entire client base from a single dashboard.
What I learned:
Domain expertise helped a lot while building. Having 8 years in strategic finance meant I never had to guess what to build, I knew exactly what SMB owners need and what they don't. That clarity saved a ton of time on product decisions.
The QuickBooks integration was very complex. OAuth, token refresh, rate limits, report parsing, account mapping, caching. It's a significant undertaking that caught me off guard early on. It was challenging without a technical background, I didn’t know what “needed” to be built and I spent a lot of time doing research making sure it was done the right way.
The main obstacle:
Two things, both outside my control on GTM.
Getting beta users before you have any social proof is brutal. Cold outreach with no app store listing, no reviews, and no existing customers to point to is an uphill battle. My email domain hit deliverability issues mid-campaign. One trial user signed up on launch day, logged in once, and never came back. It made me realize that distribution is just as hard as the product, actually it’s 10x harder.
Second, Intuit's security review. They quoted 3-10 business days. I'm now 7 weeks in. I failed the initial scan on two medium findings. Zero high or critical findings and one informational finding. I pushed fixes the same day I received the report, now I'm waiting on a rescan scheduled for June 26th. The whole security process will take 12 weeks vs. the originally quoted 3-10 business days. The entire app store GTM channel is blocked until that clears.
Where I'm going:
I should’ve been live on the app store by now, but now I’m targeting app store launch mid to late July if everything goes to this plan with cold email warming up in parallel. First goal is 10 paying customers.
Happy to talk about building on the QuickBooks API, the Anthropic integration, or navigating the Intuit review process with anyone who's interested.
This is a serious product, and honestly the Intuit delay may be forcing the right GTM question earlier than expected.
The app store will help, but I would not let it become the only launch path. For Foresight, the first 10 customers probably come less from “SMBs need FP&A” and more from a sharper pain wedge:
QuickBooks businesses that are profitable on paper but still anxious about cash, hiring, or what the next 12 weeks look like.
That is a much easier conversation than selling a full AI CFO platform upfront.
The CPA/bookkeeper tier also feels important because one firm can give you access to multiple SMBs without earning trust one owner at a time. I’d probably test both paths in parallel, but with different messaging:
For SMB owners: “know if you can hire, spend, or survive the next 12 weeks.”
For CPAs/bookkeepers: “give clients CFO-level forecasting without building spreadsheets manually.”
The cold email issue is painful, but the positioning may matter even more than volume here. A generic FP&A pitch will get ignored. A specific cash-flow/hiring-risk message to QuickBooks users has a much better chance.
Happy to put a tighter version in writing if useful. I recently did a practical GTM breakdown for an early marketplace founder, and Foresight feels like the kind of product where a focused first-10-customer plan would actually help.