I've had some version of the same conversation dozens of times in the last eighteen months. A managing partner at a large firm says something like: "We're actually thinking about building something internally. We have a few sharp people, and with Claude or ChatGPT, how hard could it be?"
I understand the instinct. The tools are genuinely impressive. The demos are convincing. And there's something appealing about the idea of owning your own technology — not depending on a vendor, not paying recurring software fees, keeping IP inside the firm.
But I've watched enough of these initiatives play out to know how the story usually goes. And I think the leaders deserve a straight answer to the question they're actually asking.
Let's take the "build" case seriously
There are real reasons to consider building. You know your firm's processes better than any software vendor does. You have control over the roadmap. You're not locked into someone else's pricing or product decisions. If you have a strong technical team and a well-defined problem, building a one-off internal tool might be the right call. I'm not here to tell you buying is always right. But audit workflow automation is not a narrow, well-defined problem. And that distinction matters enormously.
What you're actually signing up for
When a firm decides to build an AI-powered audit workflow tool, they're not commissioning a one-time project. They're committing to an ongoing engineering function that needs to be staffed, maintained, and prioritized alongside everything else the firm is managing.
General-purpose LLMs are remarkable at language. They are not pre-trained on your firm's engagement methodology or the specific way your team ties out a financial report, or the output format your clients expect. Getting a model to do that reliably across many live engagements — is the hard part.
And the work doesn't end at launch. Models update. Outputs drift. New audit standards require changes. The system needs to be monitored, validated, and maintained by someone — and that someone is your team.
"Who owns this?" is the most important question no one asks before starting a build initiative. Nine times out of ten, the answer turns out to be your best manager, spending half their time on internal tooling instead of client work.
There's also a liability dimension that doesn't get enough attention. When a purpose-built, audit-native platform produces an error, there's a vendor relationship, a support structure, and contractual accountability. When your homegrown tool produces an error in an audit file, the responsibility is entirely yours.
The "wait and see" trap is just as dangerous
Some firms aren't building. They're not buying, either. They're watching. Waiting for the technology to mature, for someone else to figure it out first.
I understand that instinct too. But I'd encourage leaders to be honest with themselves about what "wait and see" actually means in practice. It means your peer firms are pulling ahead. It means your managers are still spending hours on work that shouldn't require hours. The window to get ahead of this is real, but it's not permanent.
What "buy" actually buys you
Choosing a purpose-built platform isn't a concession. It's a decision to buy back time. For your partners, your managers — and to redirect that time toward the work that actually builds a practice.
Firms like CohnReznick didn't choose Inscope because they couldn't figure out how to use AI. They came because they recognized that building workflow automation correctly is genuinely hard, deeply domain-specific work. The question is whether your firm's energy goes into building the infrastructure or into the work that infrastructure is supposed to enable.
My take
Firms that try to build will spend months learning what the right vendor already knows. Some will ship something; most won't reach production. A few will get there and then face the ongoing cost of maintaining it. In almost every case, the total cost — in dollars, in distraction, in opportunity will exceed what a subscription would have cost.
Firms that wait will eventually buy. The only question is how much ground they'll have given up in the meantime. This isn't really a build vs. buy decision. It's a question of whether your firm leads this transition or responds to it.