Adapting with AI Part 1 of 3: The Email Every Firm Owner Dreads

The Email Every Firm Owner Dreads

It arrived on a Thursday afternoon. The client, let’s call him Marcus (not his real name), had been with us for a couple of years. He ran a small consulting firm. He asked good questions. He wasn’t a difficult client. He was an informed one. And increasingly, informed means someone who’s been reading and using AI.

His email started warmly enough, but the substance was pointed: his accounting software had been rolling out AI automation features. In his view, routine bookkeeping should be mostly automated by now. He wanted time savings reallocated to higher-value advisory work. And he wanted that to come in at or below his current budget because, in his words, AI had already commoditized the manual elements of bookkeeping.

I want to be honest about how I felt reading that.

Annoyed.

Not because the questions were unfair. Marcus was asking what a lot of clients are asking right now. The frustration was that his email reflected a version of reality the market has been aggressively selling and that my team, the one I’ve spent years building into something genuinely good, was apparently invisible within that false reality. He insinuated that we were being timid in our approach to leveraging AI in our team and in our services.

The premise buried in his message, that AI has already commoditized bookkeeping, that we should be riding a wave of automation, and that our caution is a failure of our internal AI policies, is not an actual truth. It’s a market hallucination. And it’s one every firm owner will have to confront, one client email at a time.

As a firm owner, if you’re wondering if you’re behind and won’t know how to respond to your clients, this eBook is meant to help you answer those questions.

How We Responded

My response was direct. Yes, our accounting software vendors are rolling out AI features and we’ll leverage them when we can train our team on them. But they won’t systematically cut our work in half this year. If you use generative AI tools yourself, you already know their propensity toward hallucinations. Our goal has to be fully accurate numbers for our clients, reviewed by humans. That’s not timidity, that’s the job.

The research backs this up. A McKinsey study published in November 2025 found that 88% of respondents in companies report regular AI use, but the majority are still in the experimental stages. S&P Global’s 2025 research found that 42% of companies abandoned most of their AI initiatives before reaching production, up from just 17% the year before. When it comes to accounting specifically, a joint Stanford GSB and MIT Sloan working paper found that accountants using AI closed the books 7.5 days faster each month and reallocated 8.5% of their time toward higher-value work. The technology is producing real results. But the key word is reallocation, not elimination.

The Gap This Reveals

In regard to AI adoption and usage, there are two entirely different realities running in parallel right now. In one, AI has already transformed professional services; bookkeeping is automated, prices should drop, time has been liberated. This is the reality that lives in marketing materials, tech keynotes, and increasingly, in the minds of your clients.

In the second reality, the one I hear about every week from firm owners I coach and consult with in Thriveal is that most mid-size CPA firms haven’t meaningfully deployed AI yet. Many that have tried have done so poorly. The profession is largely still waiting for software providers to hand them the next step. Accounting teams on the ground are genuinely struggling to understand how to leverage tools, both standalone and embedded in their software providers, within their specific roles. This eBook, created in collaboration with our software partner Melio, seeks to bridge that gap – and empower firms on how to effectively ramp up AI in their daily operations without going crazy. At our size, AI can’t just be something I use on my own. It needs to flow through every team member, every process, and every client interaction–most of the time without me in the room.

It’s been over three years since ChatGPT launched in November 2022, yet 2026–2027 is realistically when most firms will get this right. This year marks the beginning of a sorting: leaders, followers, and laggards. The decisions made today will determine which tier you occupy.

I don’t coach the large, well-resourced firms you read about in industry publications (which are adopting AI at higher rates than SMBs); I focus on the entrepreneurial firms of ten to fifty people, navigating this without a CTO, without an AI strategy team, often without a clear sense of where to start. What I hear is not confidence. What I hear is a version of the same question:

“Am I behind? And if I am, what do I do about it?”

This book is my answer to that question.

Every Firm Will Have This Conversation

The email Marcus sent me is not unique. A version of it is coming for every firm and may have already arrived for you. A client who attended a conference, received a pitch from a competitor promising AI-powered accounting at half the price, or who simply read one too many headlines about automation are being fed market messaging. At the writing of this eBook, we’re even seeing accounting products closing their doors that claim to be fully AI-powered, and AI FinTech tools laying off close to half their workforces. Meanwhile clients are asking: why aren’t you cutting your prices yet?

Josh Tyrangiel, writing in The Atlantic this year, documented how the CEOs of major companies and AI labs have publicly declared that AI will eliminate vast swaths of professional work. Whether or not they’re right about the timeline, their words are shaping what clients believe and how much they think they should be paying.

“The market believes AI has already transformed industries that are still in the earliest stages of real adoption.”

A Foundation Before We Begin

Before we get into what’s actually happening with AI adoption (the data, the timelines, the frameworks, etc.) there is one principle worth sharing with you:

AI is not the solution. AI is an accelerant. If your processes are strong and your team is capable, AI will amplify your results. If your processes are weak, and your team goes rogue, AI will amplify your dysfunction–faster and more expensively than you imagined.

The data backs this up. Even the Managers and the Individual Contributors in companies don’t view AI adoption the same, and this causes a bottleneck in company-wide adoption. The 2026 Small Business AI Outlook Report by business.com reported “22 percent of individual contributors view AI as having ‘anti-worker sentiment,’ compared to just 11 percent of managers. This suggests that those closest to potential job displacement are understandably more wary of the technology than those managing its implementation.”

We all need clarity about AI’s power, the fears around adoption, and the clarity of the ultimate strategy of company-wide adoption. But our profession is not quite on the same page… yet.

Everything in the chapters ahead flows from that need for clarity.

What Your Clients Believe About AI (And Why They’re Both Right and Wrong)

Marcus isn’t an outlier. He’s the leading edge of a conversation every firm is about to have with every sophisticated client. Understanding exactly what your clients believe and where those beliefs are right, where they’re wrong, and why the distinction matters is the first job.

The Commoditization Question

Your clients are not imagining things. AI is changing the economics of certain professional services work. Routine transaction categorization, basic reconciliation, common analysis of financials, standard reporting – these tasks genuinely are becoming more automatable, and that automation is real and accelerating.

But your clients have the timeline wrong. And the timeline is everything.

The research consistently shows that the gap between what AI can theoretically do and what organizations can actually absorb is wide, and crossing it takes longer than anyone wants to admit. The MIT CISR AI maturity model identifies four stages of organizational readiness, based on a survey of 721 companies:

• Stage 1 – Experiment and Educate (28% of enterprises): Educating the workforce, formulating AI policies, and experimenting with tools.
• Stage 2 – Pilot and Build Capabilities (34% of enterprises): Piloting use cases, defining metrics, and beginning to automate processes.
• Stage 3 – Develop AI Ways of Working (31% of enterprises): Scaling AI across the organization with dashboards, architecture, and a test-and-learn culture.
• Stage 4 – AI Future-Ready (7% of enterprises): AI embedded in all decision-making; proprietary capabilities developed.

The journey from Stage 1 to Stage 4 takes 24 months or more. Most small accounting firms are at Stage 1 or early Stage 2. Your clients think you’re at Stage 4.

The gap between where your clients imagine you are and where you actually are is the source of most of the pressure you’re feeling right now.

Why AI Projects Fail

S&P Global’s 2025 research found that 42% of companies abandoned most AI initiatives before reaching production–up from 17% the year prior. RAND Corporation found that AI projects fail at roughly twice the rate of traditional IT projects. The common thread: organizations treated AI as a technology problem when it was really a people and process problem. Change management, in other words–something accounting firms struggle to execute without disrupting client service.

This points to a double hallucination problem. The first is technical: AI generating confident but incorrect outputs. The second is cultural: the market believes AI has already transformed industries that are still in early adoption. The second hallucination is more dangerous, because it shapes what clients expect from your firm.

What Clients Don’t Understand About Timeline

The Stanford GSB and MIT Sloan research tells an important story. Accountants using AI tools were closing the books 7.5 days faster each month and reallocating 8.5% of their time toward higher-value work. Those are real, meaningful gains. But they’re not the overnight revolution clients imagine and they require intentional adoption, workflow redesign, and sustained effort to achieve.

Meaningful, sustainable efficiency gains from AI take 18 to 36 months of organizational work. Not 6 months. Not the quarter after you buy the software license. 18 to 36 months of intentional change, team training, workflow redesign, and quality review before the gains become reliable enough to build a service model around. The larger the team, the longer this timeline will take.

Clients think you’re in month twelve.

You’re probably in month three or even earlier.

The Confidence Gap

Recent AI research shows that only 15% of employees have received a clear AI strategy. In most firms, teams don’t know what success looks like–and if they don’t know, neither do the clients. This creates a gap where clients assume you aren’t adopting AI while simultaneously expecting the resulting savings. They will hold both conflicting beliefs until you provide a clearer narrative.

What to Take into the Conversation

When Marcus, or someone like him, sends that email, here’s the distinction that matters:

Marcus is right about the trajectory.

AI will change the economics of basic bookkeeping, and commoditization is coming for certain service lines within three to five years. But the timeline isn’t what clients imagine. As economist Austan Goolsbee observed in early 2026, the numbers don’t yet show AI affecting jobs at scale. The disruption is real–it’s just not here yet.

Marcus is wrong about the timeline. It hasn’t happened yet. Most firms are still in early-stage adoption. The gap between the marketing and the reality is real and wide.

Marcus is wrong about what efficiency means in professional services. AI cutting 30% of routine processing time doesn’t mean prices should drop 30%. It means capacity is freed to do work that wasn’t viable before.

That last point is the core of what this book is built to help you argue not defensively, but confidently, because it’s true. The chapter ahead will show you what that argument looks like in practice.

Download the full eBook

Continue reading → Part 2: Running an Efficient Team in a Market That Won’t Sit Still