The 3-phase roadmap every accounting firm needs before AI delivers

Efficiency gains from AI in accounting are no longer theoretical. But the firms that capture them share one thing in common: they built the right foundation first.

At every accounting conference I attend, the same question comes up from firm owners: “We know we need to be leveraging AI, but where do we even start?”

It’s the right question. And it’s one that Dan Astrachan, CEO of Bridgepoint Alliance, has been thinking hard about—not just for his own firm, but for every accounting practice his group looks to bring on. When we sat down recently to co-author this piece, it became clear that we were both circling the same insight from different angles.

Here's what we've both seen: the ROI from AI is measurable today. The tools are mature. The stakes are real. In fact, Karbon's 2026 State of AI in Accounting Report found that 63% of accounting professionals believe the value of a firm drops if it doesn't use AI, up 7% from the prior year. The firms capturing efficiency gains aren't chasing the flashiest AI feature, rather, they built the right foundation first.

With Dan's perspective from evaluating dozens of firms, we've mapped the journey into three distinct phases. Each builds on the last, and skipping ahead is the most reliable way to waste time and money.

Phase 1: Start with what’s already in front of you

If you're a partner at a firm between $2 million and $20 million in revenue, you probably don't have a large technology budget or a dedicated CTO. You have talented people, demanding clients, and a to-do list that never gets shorter. The idea of an 'AI transformation' sounds expensive and abstract.

It doesn't have to be. Phase 1 is about removing friction from work you're already doing. These are the table-stakes applications, and if your firm isn't using them yet, you're already behind:

• Invoice and document automation that eliminates manual data entry

• AI-assisted recruiting screening, which can cut time-to-interview by as much as 66%

• Smart timesheet suggestions that reduce the administrative burden on your team

• Automated client follow-ups with AI document categorization

• Reporting that provides actionable insights on which clients need support and which engagements are at risk

• A single source for client and internal communication, documents, and work so nothing falls through the cracks

None of these require you to rebuild your firm from the ground up. They exist in tools that already sit inside your current workflows. Every manager who spends their morning reconstructing what's happening across client engagements is experiencing a structural inefficiency. That time belongs in client work and revenue generation. Phase 1 gives it back.

Phase 2: Build the foundation that makes everything else possible

This phase is the part most firms skip. They see a compelling AI demo, buy a point solution, and wonder why the results don't match the pitch. The answer is almost always the same: fragmented data.

AI is only as good as the data you feed it. If your workflow lives in one place, your client communication in another, your time tracking in a spreadsheet, and your engagement data scattered across email threads, the intelligence layer you want cannot perform. Garbage in, garbage out is more relevant now than it has ever been.

Phase 2 is about getting your data into a single system of record: workflow, time, client engagement, and communication. For Bridgepoint, this was a non-negotiable when evaluating which accounting practice management platform to anchor their acquired firms on. A system that standardizes operations without stripping the boutique feel clients have come to expect is a rare combination. It's the reason they landed on Karbon.

Dan puts it directly: "When your firm's backbone is in a single, open system and you can bring all of your data sets together, the calculus changes."

The benefits of getting this right are immediate. Reduced administrative hours. Less time chasing clients for documents. Better visibility into what your team is actually working on. These deliver in the first months after implementation, not down the road. Karbon's AI research backs this up: firms that invest in proper AI training alongside the right infrastructure save 28% more time than those that don't, with advanced AI users saving 82 minutes per day compared to 48 minutes for beginners. The foundation and the training go together.

We've seen smart people at well-resourced firms attempt to build custom AI tooling in-house, only to discover that the cost of security review, ongoing maintenance, and compliance oversight quickly outpaces any projected savings. Your clients' data is so sensitive that the infrastructure required to protect it is not a side project.

Dan shares his direct experience with this:

"At Bridgepoint, we were comfortable building internal tools, especially as platforms like ChatGPT and Claude lowered the barrier to development. It was exciting to see our team shipping real solutions with meaningful time savings.

But the dynamic changed when those tools started touching confidential client data. What worked internally didn't translate to the level of enterprise-grade security required for our most sensitive information.

Working through it with our CTO, we realized that owning these tools, along with the ongoing security reviews and operational burden, would ultimately be more expensive than the value they created. We're not a software company.

So we stepped back, reviewed our core value proposition to both our clients and our professionals, and decided to partner with leading software vendors, which gave us back peace of mind."

Phase 3: The intelligence layer, and why it changes everything

Once you have clean, centralized data and your team is actually using it consistently, you've unlocked something most firms haven't even imagined yet.

The intelligence layer is where accounting firms start operating like large advisory practices, without the headcount. As Dan puts it, some of what becomes possible includes:

• Early warning signals on client issues, surfaced before they become problems

• Firmographic analysis that tells you which services to offer to which clients before the client thinks to ask

• Predictive staffing models that flag burnout indicators in your team before someone walks out the door

• Proactive cross-sell signals that let you grow revenue per client without adding headcount

Consider that the average small accounting firm serves its clients 1.2–1.4 services, with the Big 4 averaging 6–8. This difference is about capacity and visibility, not talent. With the right data infrastructure in place, firms can finally see the full picture of each client relationship and act on it.

From Dan's perspective:

"Being partners with a solution like Karbon, a platform that brings together everything from full client intelligence, to project status and employee utilization, that's the context window you need to have in place to deliver the true benefits of AI. Karbon is in a unique position because it connects all of these tools through an open platform architecture, which expands the depth and quality of context available across the firm."

What about the soul of the firm?

Many firm owners who hear the words "AI automation" ask the same underlying question: what about the people? What happens to the relationships we've spent decades building? What happens to the team and the business that I have invested so much in?

Those are the right questions to ask. Automation, done well, protects exactly what makes your firm valuable to clients.

I have a true devotion to this profession and love to hear small businesses sharing the impact their accountant has had on their business and personal lives. I have a favorite question I ask every business owner I encounter. Whether at the local meat store, the dentist, or the boutique, I ask them who their accountant is and what's the biggest thing they've done for them. 

Almost every time, the answer is a story about a trusted relationship, not a transaction. The accountant is the number one trusted advisor to a small business. That's not going to change because of AI.

What AI does is give accountants more time to be that advisor. Less time chasing documents. Less time reconstructing status updates. Less time on rote tasks that don't require your expertise. More time on the conversations that actually matter. The data supports this: 82% of accounting professionals say AI has had a positive impact on their firm, with client relationships and communication ranking among the top benefits.

And more services to offer. Not just bookkeeping, tax, and audit—but HR services, technology advisory, lending referrals grounded in an intimate understanding of a client's financial health. Firms that build this kind of trusted, full-service relationship with small businesses will pull away from the competition over the next three to five years.

The sequence matters, and so does getting started

The firm leaders who will look back on this period with confidence are the ones who followed the sequence. They did the foundational work first. They chose platforms built to handle the security and compliance burden so they didn't have to. And they moved while the window was still open.

The firms acting now have a real head start. That gap compounds every year.

A $500 million transformation budget is not the requirement. A clear sequence, the right system of record, and the organizational commitment to embed it into how your team actually works—that's what separates the firms that capture the value from the ones that checked the box.