Where Does Your Firm Sit on the AI Maturity Curve?

Most accounting firms are using AI. The important question is whether they're actually getting value from it.

There's a meaningful difference between a firm that has staff occasionally pasting things into ChatGPT, and a firm that has restructured its processes around AI infrastructure. The gap in business value between those two positions is enormous, and it's widening every month. Understanding where your firm sits, and what the path forward looks like, is one of the most important strategic conversations you can have right now.

At AI Dojo, we've mapped this progression into what we call the AI Maturity Curve: a four-stage model that describes how accounting firms evolve from tentative AI experimentation to genuinely AI-driven operations. It's not a theory, it's what we're seeing play out across the firms we work with.

Stage 1: AI-Supported

Most firms begin here. Staff are using AI tools such as ChatGPT, Microsoft Copilot, or Claude for drafting emails, summarising documents, or doing quick research. It's genuinely useful, and the time savings are beneficial.

However, the value of AI at this stage is capped by the individual. How much benefit a team member gets from AI depends almost entirely on their personal prompting skill and willingness to experiment. There's no firm-level leverage. There's no consistency, no institutional knowledge being built, and no compounding effect across the practice.

If your firm's AI story is "we let people use ChatGPT", you're in Stage 1. And while it's a fine place to start, it's not a competitive position you want to stay in for long.

Stage 2: AI-Enhanced

The leap from Stage 1 to Stage 2 is where things get interesting, and where tangible, measurable ROI starts to appear.

At this stage, AI is no longer a general-purpose tool that staff use ad hoc. It's embedded directly into the firm's core processes. Examples are purpose-built AI assistants for FBT workpapers, BAS preparation, month-end reconciliations, payroll reviews. Standardised inputs, consistent outputs, repeatable processes.

The numbers speak for themselves. A firm processing 20 FBT returns per year, where each one drops from four hours to two hours, frees up roughly 40 hours annually from a single process. Multiply that across BAS, tax planning, year-end compliance, and reconciliations, and you're talking about a meaningful recovery of capacity that can be redirected to client advisory work.

The limitation at Stage 2 is that the skills still exist in silos. Each process runs in isolation, and there's duplication of logic across different tools and file types. It's efficient, but it's not yet integrated. The firm is smarter, but the AI isn't yet learning at a firm level.

Stage 3: AI-Enabled

Stage 3 represents a fundamental shift in how the firm is structured. Rather than AI being a tool that staff reach for, the firm itself is built around AI infrastructure. This is the point at which the competitive advantage becomes difficult to replicate quickly.

At this stage, firms are running connected AI workspaces with client memory, Xero integrations, custom skill libraries, and admin dashboards that give partners visibility across how AI is being used and performing. Critically, feedback loops begin to compound accuracy. Every interaction makes the system smarter, not just for one user, but across the whole firm.

The value driver here isn't just efficiency, it's advisory capacity. When compliance work is handled faster and more consistently by AI, accountants have the bandwidth to have more meaningful conversations with clients. We've seen firms at this stage with approximately 60% of their revenue being touched by AI in some way. Not replaced by it, but genuinely supported and accelerated by it.

The main limitation to acknowledge is that at Stage 3, AI still waits to be asked. Humans initiate processes, and AI executes them. That's not a criticism, it's appropriate for where governance and trust are at this stage. But it points to what's coming.

Stage 4: AI-Driven

Stage 4 is where the curve inflects sharply upward, and where the business model of accounting fundamentally changes.

In an AI-driven firm, processes are triggered automatically. A client document arrives, and the tax process initiates itself. AI executes the work. Humans review exceptions. The firm doesn't need more people to handle more volume. It scales on platform, not headcount.

This is what agentic AI means in practice, and it's no longer a distant concept. The firms building toward this now are the ones who will define what modern accounting looks like in five years. But it comes with real requirements: mature governance frameworks, clear audit trails, and robust exception-handling protocols. This isn't something you shortcut your way into, and any firm claiming to be here without those foundations is taking on risk they may not fully appreciate.

A Word on Culture

Reading through these four stages, it's tempting to think the progression is about software. It isn't. It's about organisational posture, how leadership thinks about AI, how processes are designed, how staff are trained, and how seriously the firm is willing to commit to systematic change rather than ad hoc adoption.

The firms that move fastest through this curve aren't necessarily the largest or the most tech-savvy. They're the ones where leadership has made a deliberate decision to treat AI as infrastructure, not as a productivity experiment.

As you read this, it's worth asking yourself honestly: which stage are we at? Not which stage do we aspire to, or which stage makes us look good in a conversation, but where are we actually operating today?

The gap between Stage 1 and Stage 3 is growing faster than most firms realise.