The Journal of Accountancy’s May 2026 cover story identified five competencies CPAs need as AI takes over routine analytical work: critical thinking, creativity, empathy, strategic vision, and interpersonal relationships. The premise is straightforward. AI handles the pattern recognition. Humans handle the judgment.

That framing is correct, but it skips a step. Between “AI handles the data” and “humans provide the judgment” sits a technology layer that most CPA firms haven’t thought through.

The competencies assume your tech works

Take critical thinking. The article describes a scenario where an AI tool flags a variance in gross margin, and a CPA has to determine whether it’s a real risk or a normal fluctuation. For that to work, the AI needs access to accurate, complete data. Which means your document management system is organized. Your integrations are pulling the right records. Your team knows which data sources feed into the AI’s analysis.

If your data is scattered across personal OneDrive folders, email attachments, and a shared drive nobody maintains, the AI output your CPAs are supposed to “think critically” about is garbage. Critical thinking applied to bad data still produces bad conclusions.

Or take the interpersonal relationship piece. The article suggests CPAs should shift from transactional interactions to relational ones, asking clients “what’s been most challenging for you lately?” instead of just reviewing financials. That’s good advice. But if your team is spending three hours troubleshooting a printer, waiting on IT tickets, or manually reconciling data between systems that should talk to each other, there’s no time left for relational work. The administrative burden eats the advisory capacity.

What the IT stack actually needs to look like

If you accept that AI handles data analysis and your people handle judgment and relationships, the technology layer between them needs to do a few things:

Give AI clean data to work with. That means a document management system with consistent naming, folder structures, and permissions. It means integrations between your practice management software, your accounting platform, and your client portals that actually sync. It means someone has audited what Copilot or whatever AI tool you use can access, and confirmed it’s seeing the right information.

Keep client data where it belongs. If your staff is using AI tools to draft memos, analyze financials, or summarize client situations, where does that data go? Is it staying within your firm’s tenant, or is it being processed by a consumer AI tool with no data retention guarantees? An acceptable use policy needs to answer this. Which tools are approved, which data can go into them, what’s off-limits.

Reduce the time your people spend on things that aren’t advisory work. Every hour a CPA spends fighting technology is an hour they’re not building client relationships or applying professional judgment. And those hours add up fast. Beyond the usual IT headaches, tax preparers now have to maintain a Written Information Security Plan under IRS requirements and the FTC Safeguards Rule. That’s another compliance obligation landing on people whose job is supposed to be advisory work, not security documentation. This is the basic argument for managed IT: keep the infrastructure running and the compliance boxes checked so your team can do the work they’re trained for. AI makes it more urgent because the gap between “firm that works efficiently” and “firm that doesn’t” is about to widen considerably.

Make AI tools governable. If your firm is evaluating Copilot-bundled M365 licenses (new SKUs available July 1), someone needs to configure what Copilot can access before you buy in. That means data classification so sensitive client information doesn’t surface in the wrong context, and a clear definition of what “review before sending” means when AI drafted the first version.

The firm-level question

The Journal of Accountancy article is aimed at individual CPAs: develop these competencies, stay relevant, don’t get replaced. That’s fine for personal career advice. But at the firm level, the question is different: does your infrastructure support a team that works this way?

Most small CPA firms we talk to are somewhere in the middle. They have M365 but haven’t configured it beyond the basics. They have a document management system but it’s half-adopted. They’re curious about AI but haven’t written a policy about it. Their team is capable of advisory work but drowning in administrative overhead.

The five competencies the Journal identified are real. But they don’t happen in a vacuum. The firms that actually make this shift will be the ones whose technology gets out of the way and lets people do the work AI can’t.

Artech Solutions works with CPA firms and professional services companies in Iowa to build IT environments that support modern workflows. If your firm is thinking about AI readiness, let’s talk.