By Biju Samuel, CIO, Frazier & Deeter.
If you work in a public accounting firm today, you’re getting pitched “modernization” from every angle: AI tools, automation platforms, new billing systems, workflow engines and more.
It’s tempting to solve every pain point by buying another too, but if you’re not careful, you end up with what I call a Frankenstein tech stack—a patchwork of systems stitched together with manual workarounds. Your people become “middleware” not because they should but because the systems leave them no choice.
At Frazier & Deeter, we decided to take a different approach. Instead of starting with software, we started with the process. We mapped how work flows across the firm, from the first conversation with a prospect all the way to collecting cash, and then designed technology to support those workflows.
This article shares what that looked like in practice, how we made build vs. buy decisions and what we’ve learned about avoiding disconnected systems that slow down firms.
Recommended Articles
What a “Frankenstein Tech Stack” Looks Like
When I say “Frankenstein tech stack,” I’m talking about this pattern: You use one provider for one part of the process, another provider for something adjacent and then you rely on people to move data back and forth between them.
A very real example in our world:
- Our tax practice tools contained important data and checklists.
- Our workpapers and related files lived in CaseWare.
- The “integration layer” between those systems was a human being manually copying and pasting information.
That’s a Frankenstein stack. You have multiple logins, no real integration and manual transfer of data. For every copy‑paste, even with a great team like ours, there’s a chance for an error. You see slower turnaround on client engagements and frustration in the staff who feel like the tools are decades old—because in many cases, their underlying design is.
So the question becomes: How do you modernize without building the monster?
My answer: process before platform.
Why We Started With Prospect‑to‑Cash: The 90+ Hour Workflow Mapping
Like many firms in legacy industries, when I joined Frazier & Deeter, our technology function was largely in firefighting mode, responding to tickets instead of being seen as a strategic partner to the firm’s greater purpose. That’s a very common starting point for IT in professional services firms. Leadership made a deliberate decision to change how we function. My team decided: before we buy or build anything new, we’re going to learn the business in detail.
We focused on the prospect‑to‑cash journey because that’s the spine of the firm’s economics. Rather than ask, “Which CRM should we buy?” or “What’s the best billing tool?” we asked: “How does the work actually flow today, step by step, and where are the opportunities for improvement?”
To find out the answer, we ran more than 90 hours of structured workflow sessions across the firm. Each session was about four hours. The meetings consisted of a whiteboard, sticky notes and us hearing at every level how clients journey through the firm. Once we had everything on the wall, we translated it into Visio and Lucidchart diagrams so the flows were easier to see and share.
That became the foundation for every technology decision that followed.
How We Evaluated Build vs. Buy
One of the most important decisions in any modernization effort is whether to build, buy or some mix of the two. Because we had done the job of workflow mapping, we had the data needed to make those calls more intelligently.
When We Decided to Build (or Custom‑Integrate)
We leaned toward building custom solutions when the process was core to how we deliver value, the way we worked was meaningfully different from what generic tools assumed and the cost of forcing the business into a generic workflow was higher than the cost of building. We also do it when it’s easy and when the ROI of automating is a clear benefit to time, quality and experience.
The “last mile of tax” integration is a good example of where we chose to build rather than buy:
At the beginning of this article, I mentioned an example of how the FD team was manually copying and pasting information from one system to another. This was happening at the “last mile” of our tax process. In response, we built an integration that automatically pulls the required data from our PPC tools and feeds it directly into CaseWare. That one automation removed a repetitive copy‑paste step, reduced the risk of manual errors and gave our tax professionals back time for review, judgment and client-facing advisory.
When We Decided to Buy
On the other hand, our workflow maps made it obvious that we were missing some foundational platforms, and some things made sense for us to bring in rather than do ourselves. Think: a modern CRM to manage client and prospect relationships and a robust professional services automation (PSA) layer to tie together projects, resources and financials. Those are not niche, firm‑specific problems. They’re core infrastructure problems, and there are mature platforms that exist to solve them.
In other words, we chose to buy platforms where the capability is widely needed and commoditized—then build firm‑specific automations on top of those platforms.
Modernization Without the Monster
Modernization isn’t about how many tools you buy or how prominently AI features in your roadmap. Successful tech stacks, whether built or bought, rely on integration and design, not the origin of the tools. Both custom and off-the-shelf components cause issues when disconnected from existing systems. The key is that every component fits the end-to-end workflow, shares data cleanly and reduces manual work. A well-integrated mix is the goal; a pile of unconnected tools creates the “monster.”
If there’s one lesson from our journey at Frazier & Deeter, it’s this: Put process before platform.
Map how work really flows, end to end. Invite people at every level into that conversation. Use what you learn to decide where to build, where to buy and where to integrate thoughtfully so your systems act like a coherent whole instead of a stitched‑together monster.
When you do that, IT stops being a ticket desk and becomes a strategic partner that helps your firm grow, serve clients better and give your people the modern experience they expect.
—
Biju Samuel is Chief Information Officer at Frazier & Deeter, where he leads technology, automation and AI strategy for one of the nation’s fastest-growing accounting and advisory firms. With 25+ years of experience across professional services and software-driven industries, Biju is a strong advocate for responsible technology adoption, with a focus on data privacy, governance and human-in-the-loop design. He regularly advises leadership on how automation can free professionals to focus on higher-value, complex client work without compromising security or professional judgment.
Thanks for reading CPA Practice Advisor!
Subscribe Already registered? Log In
Need more information? Read the FAQs
