Evaluating which solutions your firm does and does not need has never been easy, but in these days of bot-written reviews and paid influencers fueled by the AI-for-everything boom, it has become even more difficult to separate what's real from what's hype. Tech-forward CPA firms have found that effective vendor vetting requires a systemic, intentional approach grounded not in wishful thinking but concrete business needs. Virginia-based YHB expressed this approach in terms of problem-solving.
"Our process starts with the business problem, not the product," said Jeremy Shen, YHB's chief strategy officer. "If a tool is not clearly tied to a specific workflow improvement or client outcome, we do not pursue it. From there, we prioritize real-world validation over polished demonstrations. We test solutions against actual workflows, with real data and real users."
Taking a concrete approach means not taking claims at face value. Tucker Haas, chief technology officer for Top 50 firm Crete Professionals Alliance, was one of many people who emphasized trying the tool for yourself outside the vendor's control and running stringent tests.
"We separate hype from reality by running pilots," he said. "We do not commit based on demos. We pick a firm, deploy the tool in a controlled environment, measure outcomes against specific KPIs, and then decide whether to scale. If a vendor claims 50% time savings, we validate that with our own data before recommending it to other firms."
Sam Leon, head of Millennial CPA, a sole proprietorship based in Virginia, noted that hands-on testing is especially important for anything AI-based, considering the confusion and ambiguity that often surrounds such solutions.
"Always get at least a short period of free demos. Test them against paid ChatGPT/Claude to discover if there's an actual new use case, or is this just AI in a wrapper? That's not to say pieces of software can't involve AI-wrapped tools, but it must provide a new value that LLMs and agents alone can't provide," he said.
While demos can certainly look impressive (after all, it's what they're designed to do), it's important not to lose sight of the mundane realities of buying and maintaining a solution. Scott Sanders, chief information officer of Top 50 firm Sikich, said his firm thinks less in terms of how impressive the pitch was and more in terms of how the purchase will affect things in the long term, both positively and negatively.
"Substance over spectacle drives every evaluation," he said. "In an era where a compelling demo can be assembled in hours, we look past the pitch to what matters: total cost of ownership, realistic timeline to value and how a solution integrates within our existing ecosystem and processes. If the business case doesn't hold up, the conversation ends there."
This leads to another aspect that several firms mentioned: integration. Does a solution play well with the rest of the firm's tech stack? Considering how often automated workflows require several different solutions working together, many firm executives said that if something cannot interact with the rest of their technology, it's not worth buying. Mike DeKock, head of Iowa-based MJD Advisors, was quite blunt on this point.
"One requirement is non-negotiable. Tools need to be API-first and able to integrate cleanly into our workflows. If a system can't connect to our data and operate as part of a broader system, it's a non-starter. We're not interested in adding disconnected tools that create more work," he said.
Shen, from YHB, raised the same point, as without integration there is little value to be had in the firm's broader tech stack, which needs to act in concert to reach true efficiency.
"Integration is a critical filter," he said. "If a product cannot fit into our broader ecosystem or requires excessive workarounds, it does not move forward regardless of how strong it looks in isolation."
However, beyond the systemic evaluation, it's also important to trust your instincts. DeKock noted that his firm relies heavily on taste, and he evaluates new products personally to see how well they fit into, and improve, their current systems.
"That perspective shapes a lot of our decisions. We're less concerned with feature lists and more focused on whether something is well-designed, intuitive and aligned with how we work," he said.
Who is in the discussion can be as important as what is discussed in the first place, Sanders pointed out. Sikich's selection process is broad, incorporating many people from different parts of the firm so it accounts for the firm-wide impact of a tech decision versus just one part.
"Sikich approaches vendor evaluation with the same rigor we apply to client engagements," said Sanders. "Every discussion, from initial review through pilot to potential onboarding, includes members of our internal innovation teams who bring deep industry expertise to the table. This ensures that technology decisions are never made in isolation from the people who understand our business and our clients best."
If, after an exhaustive search, a firm still hasn't found the solution that fits its needs, leaders do not necessarily have to compromise with one that's close enough; theoretically, they could also build their own, as Crete sometimes does.
"For categories where no vendor meets our bar, we build," said Haas. "That is how our TaxAI product came to be. We evaluated every solution in the market, including legacy incumbents like SurePrep, and concluded we could deliver something fundamentally better. But building is always the last resort, not the default. We would rather partner when we can and focus our engineering on the areas where we have a genuine edge."
