Everywhere you turn right now, the message is the same: “You need to be using AI.” Clients are asking about it. Conference agendas are full of it. Vendors are promising it. And yet, in quiet conversations with leaders, I keep hearing something very different:
“We know we need to do something…”
“We’ve dabbled, but it feels scattered.”
“Our team is experimenting, but we don’t have a strategy.”
“I’m not even sure where to start.”
That tension is what I call the AI adoption gap. It’s the space between knowing you should and knowing how. And that gap is especially real in accounting firms. Because when your business is built on precision, trust, regulation, and client relationships, experimentation feels riskier. So let’s make this practical. Here’s what I’ve learned about closing the AI adoption gap, without creating chaos.
One of the biggest mistakes I see is this: Leaders assume AI adoption requires a sweeping transformation. It doesn’t. It requires a starting point.
“How do we use AI across the firm?”
“Where does time get burned in one repeatable workflow?”
Then choose a wedge:
Create – drafting emails, proposals, client updates, performance reviews
Condense – summarizing long email threads, tax notices, research, meeting notes
Coordinate – capturing action items, generating follow-ups, organizing next steps
That’s it. For an accounting firm, this might look like:
Drafting first-pass client communications during busy season
Summarizing IRS updates into internal briefings
Turning partner meetings into clear action plans
Creating first drafts of SOPs after a merger
You don’t need an AI strategy deck to start. You need one workflow. One wedge. One repeatable improvement. Repeatable beats impressive every time.
Another source of confusion is that we lump everything under “AI.” But there’s a meaningful distinction:
This is where most firms should start. It helps you:
Draft
Summarize
Generate options
Improve clarity
You remain fully in control.
For example:
Draft a first-pass client update on new tax legislation.
Summarize a complex client email into key issues and next steps.
Generate three approaches to restructuring a CAS engagement.
You review. You decide. You execute. That alone can save hours per week.
This is more advanced. Instead of responding to a prompt, AI operates within a loop:
Observe (gather inputs)
Interpret (extract meaning)
Reason (decide next best action)
Act (draft, route, notify)
Evaluate (improve over time)
In an accounting firm, that might eventually look like:
Intake → triage → draft response suggestions
Automated follow-ups on missing client documentation
Routing requests to the right service line based on content
But here’s the key: Don’t build agents before you build habits. Master traditional AI first. Build comfort. Establish guardrails. Then expand.
One of the smartest things any firm can do is establish simple, clear principles before scaling AI use. Because here’s the truth: AI accelerates work. It can also accelerate mistakes. Especially in accounting firms, where data sensitivity and regulatory risk are real.
Basic guardrails might include:
Never input client information into public tools.
Always require human approval before sending external communications.
Clearly define what data is off-limits.
Validate outputs before relying on them.
The question I like to use is simple: “Would I be comfortable if this input or output appeared in a security audit?” If the answer is no, stop. Speed without guardrails is just faster risk.
The firms that will win with AI won’t be the ones with the flashiest announcements. They’ll be the ones who:
Encourage small experiments.
Share what’s working across teams.
Pilot before scaling.
Create forums for ideas to be reviewed and refined.
Innovation needs structure to scale. Especially in accounting firms, where independence across offices or partners can create fragmentation, shared governance matters. Not to slow people down. But to prevent 12 different tools solving the same problem in 12 different ways.
If you’re a leader in an accounting firm, here’s your simple starting point:
Pick one recurring workflow.
Identify where time gets burned.
Choose one wedge (Create / Condense / Coordinate).
Test one safe, repeatable improvement next week.
Not next quarter, Next week. Then do it again. The AI adoption gap doesn’t close through urgency. It closes through repetition.
And the firms that build repeatable discipline now will look back in three years and realize they didn’t just “adopt AI.” They redesigned how work flows.
That’s the real opportunity.