Build vs Buy

The Firm OS is in place. The second brain is capturing, structuring, and surfacing intelligence. The signal infrastructure is running. The policies are set.

Now comes the question every firm owner eventually arrives at, usually after watching a vendor demo that made everything look effortless.

What do we actually need to buy?

It is the wrong question. Not because buying is wrong. Because buying before thinking about what problem you are solving produces technology debt faster than almost any other decision a firm can make. The tool that seemed essential in the demo becomes the tool nobody uses six months later, sitting in the stack, drawing a monthly fee, and making the practice manager feel vaguely guilty every time the invoice arrives.

Part 5 is about making better technology decisions. Not which tools to buy. How to think before you buy anything.

The accounting technology market is unusually noisy.

Every category has multiple vendors. Every vendor has a conference presence, a case study library, and a partner programme that puts their logo on the marketing materials of the firms that use them. The signals that used to tell a firm which tools were actually working, peer recommendations, independent reviews, word of mouth, are increasingly hard to separate from the vendor-funded content that surrounds them.

Jason Staats identified the most reliable signal in the accounting tech market from his dataset of nearly 1,000 firms. Peer experience drives adoption more than vendor marketing. Firm owners talking to other firm owners. Hearing directly what is working and what was a mistake. That signal is what actually changes firm decisions.

The challenge is finding it. The accounting technology ecosystem has become sophisticated enough that even peer recommendations can be shaped by referral programmes, affiliate relationships, and the natural enthusiasm people feel for tools they have recently adopted and are still in the honeymoon period with.

The intelligent firm does not buy on enthusiasm. It buys on evidence. And it only goes looking for evidence once it has been clear about what problem it is trying to solve.

The build versus buy question has become more complicated than it used to be.

Until recently the answer was almost always buy. Building custom software required developers, timelines, and budgets that most accounting firms could not justify. The tools available off the shelf were imperfect but they were better than what a firm could build for itself.

That calculation has shifted.

Anderson Petergeorge is a CPA who built his career across KPMG, Scotiabank investment banking, and five years at CPP Investments doing private equity before working with the UNDP as a financial consultant helping businesses in Colombia, Guatemala, Micronesia, Sri Lanka, and Uganda attract investment. He then came back to the accounting profession and served forty clients as a solo practitioner before co-founding Quanto, a technology platform backed by a16z Speedrun and Google for Startups that automates sales and onboarding for bookkeeping firms using AI agents.

Anderson built Quanto because he experienced the problem firsthand. As a solo accountant serving forty clients, he was doing the same manual work over and over that no off-the-shelf tool was solving exactly the way he needed. So he built it himself.

At a live workshop with Brian Clare from Blueprint Accounting, Anderson demonstrated what this looks like in practice for any firm willing to try. They built a working pricing calculator in under thirty minutes using ChatGPT and a no-code tool. No engineering team. No custom development project. No complicated tech stack. The pricing logic the firm needed, built by accountants for accountants, in an afternoon.

The implication is not that every firm should build everything itself. It is that the line between what requires a vendor and what a firm can build internally has moved significantly. Firms that do not know where that line sits are either paying for tools they could build themselves or avoiding tools that would genuinely help them because they assume it requires resources they do not have.

By the end of Part 5, the firm will have a framework for making technology decisions with clarity rather than under pressure. It will know which questions to ask before any purchase, how to evaluate vendors without getting sold a roadmap that never arrives, and what the intelligent firm's technology stack actually looks like when all the choices have been made well.

The Firm OS is in place. The next decision is what to connect it to.