The most useful AI conversation for professional-services firms is not "Which model should we buy?" It is "Which workflow should we improve, how will we measure it, and who owns the final output?" Recent articles from CPA Practice Advisor and Artificial Lawyer make that point from different angles: accounting firm leaders need better measurement systems, while the legal market is moving toward more vertical AI tools and higher-abstraction work.
Start with the metric, not the AI tool
CPA Practice Advisor's column on metrics describes a familiar firm-owner problem: tracking the numbers that fall out of the accounting system, such as gross profit margin, net income, and revenue, without enough visibility into what is actually driving performance.
In the article, the firm owner could see gross profit margin declining but could not explain why. The open questions were operational ones: staffing, pricing, and efficiency. That is exactly where custom AI and automation work should begin for accounting firms and other professional-services firms.
A useful workflow target is not simply "add AI." It is a specific operational question: where are we overstaffed, underpriced, slow to respond, or losing margin? If the firm cannot see the cause of a metric, automation may only make the blind spot move faster.
Vertical legal AI raises the integration question
Artificial Lawyer reports that OpenAI is planning a legal AI offering that could be branded as "Codex for Legal," joining Anthropic and Microsoft in providing legal-specific tools for lawyers. The same report notes that legal AI tools could take the shape of plugins, and that Claude for Legal has expanded to plugins and MCP connectors to legal tech software.
For law firm leaders, this is a sign that legal AI is moving from general chat interfaces toward more specialized tools connected to firm systems. That makes workflow design more important, not less important.
The commercial question becomes: which legal process should be connected, what source systems does it depend on, and what review step protects professional judgment? A vertical AI tool may help with the work, but the firm still has to define the workflow around it.
Professional ownership does not disappear
In Artificial Lawyer's interview with Cooley's innovation chief David Wang, one theme is that legal writing is evolving toward higher abstraction levels, with AI enhancing productivity while requiring firms to redefine professional ownership of the final output.
That point matters for both law and accounting firms. Agentic workflows can draft, summarize, route, compare, or prepare work for review. But the firm still needs a clear rule for who is accountable for the client-facing result.
This is also a training issue. The same interview notes anxiety among junior lawyers as roles shift toward productivity maximization, and frames AI as a skill development tool. Firms adopting automation should make ownership and development explicit rather than treating AI as a background utility.
Use a sober AI lens: measure before scaling
The White Box's recent AI analysis describes unusually steep investment in AI while also cautioning that, when compared with global GDP, the situation is not unique in history. That is a useful operating posture for firm owners: take AI seriously, but do not let market noise set the roadmap.
The safest first AI workflow is usually the one tied to a visible business problem and a measurable operating metric. For an accounting firm, that may be a client intake, pricing, staffing, or margin visibility workflow. For a law firm, it may be a drafting, review, intake, knowledge retrieval, or matter-management workflow with a clear human review layer.
The common pattern is the same: pick one workflow, connect it to the right systems, define the review step, and measure whether it improves the metric that matters.
- Do not start with a generic AI tool; start with the metric or workflow problem the firm cannot currently explain.
- Vertical legal AI tools make integration, review, and ownership decisions more important for law firms.
- Accounting firm automation should be tied to margin, staffing, pricing, or efficiency visibility, not just back-office novelty.
- Junior professional development and final-output accountability should be designed into AI workflows from the beginning.
Sources watched
- Column: The Day I Got Serious About Metrics (CPA Practice Advisor AI)
- David Wang, Cooley - Legal Innovators California Keynote (Artificial Lawyer)
- OpenAI Plans 'Codex For Legal' (Artificial Lawyer)
- My Honest & Sober Opinion on AI (The White Box)
