Setting AI Goals That Actually Mean Something

Why budget follows strategy — and how defining success changes everything

Here’s a number worth sitting with: almost nine out of ten organizations now use artificial intelligence regularly. And yet only 6% are actually capturing meaningful enterprise value from it.

That gap — between AI adoption and AI transformation — has never been wider. And for accounting firms, it’s even more pronounced. Most firms are somewhere in the middle: past curiosity, not yet at impact. They’re using AI. They’re just not sure what they’re trying to accomplish with it.

That’s the problem this article addresses. Before you can measure success, you have to define it. And before you can define it, you have to understand why most firms never do.

What Actually Drives AI Adoption — And Why It Matters

Research consistently shows that the motivations driving AI adoption across industries are not what most people would expect. According to a 2025 G2 survey of professionals across industries, nearly 40% of companies cite operational efficiency as their top reason for investing in AI.

Another 27% point to product innovation.

And

20% say they’re doing it simply to stay competitive.

What’s striking about that last number is what it really means: one in five companies is adopting AI primarily because their competitors are. Not because they’ve identified a specific problem. Not because they’ve defined a target outcome. Because they don’t want to be left behind.

UK government research published in January 2026 put it plainly: businesses not yet using AI tend to adopt reactively, driven by external pressures such as client demands or industry trends rather than internal strategic vision.

In other words, FOMO — fear of missing out — is a primary driver of enterprise AI adoption. And reactive adoption, as the data makes clear, almost never produces lasting results.

How Accounting Firms Compare

Accounting firms follow the same reactive pattern — and in some ways, they’re more susceptible to it than most industries.

The profession is built on precision, compliance, and risk management. Those values don’t naturally encourage bold experimentation. So when AI began generating real buzz, many firms didn’t respond with a strategic plan. They responded with caution, then with urgency, then with action — in that order, and usually without a clear destination.

92% of accounting firm leaders say they’re familiar with AI — but only 14% consider themselves extremely familiar. Source: BILL / NewtonX survey of 207 accounting firm leaders, 2026

A 2026 survey of 207 accounting firm leaders found that while awareness is nearly universal, only 14% consider themselves extremely familiar with AI applications in accounting. Most firms are still deciding how far and how fast to push it — which means most firms are still reacting to the market rather than leading with a strategy.

The numbers on outcomes are sobering. More than half of accounting firms are still in the experimentation phase. And according to MIT’s State of AI in Business 2025 study, 95% of generative AI pilots fail. Not 15%. Ninety-five.

Enterprises without a formal AI strategy report only 37% success in AI adoption. Those with a strategy report 80%. Source: Writer.com Enterprise AI Adoption Report, 2025

That single data point — 37% versus 80% — is the clearest argument for why strategy has to come before tools, before budgets, and before timelines. A formal AI strategy more than doubles your odds of success. And the foundation of any real strategy is a clear definition of what success looks like.

The Size Question: It’s Not One-Size-Fits-All

Before we talk about how to set goals, it’s worth acknowledging... firm size changes everything about how AI strategy is built and resourced.

Research from the OECD confirms what practitioners already know — large firms are roughly twice as likely to adopt AI as medium-sized firms, and several times more likely than small enterprises. The gap isn’t about interest. It’s about infrastructure, budget, and capacity to absorb change.

A 50-person regional CPA firm and a 5-person boutique practice are not running the same AI playbook. Nor should they. The goals, the investment levels, the timelines, and the definitions of success will look completely different — and that’s appropriate. What can’t differ is the discipline of defining those things intentionally before you start spending.

For Smaller Firms

The temptation for smaller firms is to either over-invest in technology they don’t have the capacity to support, or to under-invest and fall further behind. The right answer is neither. It’s to identify one or two high-impact use cases, resource them properly, and measure results before expanding. A small firm that successfully automates one workflow and can prove the ROI is better positioned than a larger firm that has ten tools nobody uses consistently.

For Mid-Size and Larger Firms

Larger firms have more capacity — but also more complexity. More stakeholders. More legacy systems. More workflows to consider. For these firms, goal-setting needs to happen at multiple levels: firm-wide strategic goals, practice-area goals, and individual workflow goals. Without that structure, AI initiatives compete for the same budget and attention and nobody wins.

Innovation Requires Dedicated Resources — Not Borrowed Ones

Here’s one of the most common mistakes firms make when budgeting for AI... they treat it as an add-on to someone’s existing role.

A senior manager gets tasked with “overseeing the AI initiative” alongside their full client load. A partner approves a tool purchase and assumes the team will figure it out. An associate is asked to “champion AI adoption” in their spare time.

This approach virtually guarantees mediocre results. Not because the people aren’t capable, but because innovation — real innovation, the kind that changes how a firm operates — requires focused attention. It competes poorly with billable hours. It gets deprioritized when client deadlines loom. It stalls, not from lack of interest, but from lack of protected time and resources.

McKinsey’s 2025 research is instructive here: AI high performers are three times more likely than their peers to have senior leaders who not only support AI initiatives but actively demonstrate ownership of them. Not delegating. Owning. That distinction matters enormously.

If AI is genuinely a strategic priority for your firm — and the data suggests it should be — it needs to be resourced like one. That doesn’t necessarily mean a dedicated hire, especially for smaller firms. But it does mean protected time, clear accountability, and a budget line that isn’t the first thing cut when a busy season hits.

AI high performers are 3x more likely to have senior leaders who demonstrate active ownership — not just support — of AI initiatives. Source: McKinsey State of AI, 2025

Should Budget Drive Your Goals, or Should Your Goals Drive Your Budget?

Most firms approach AI investment the way they approach most technology spending: they set a budget first, then figure out what to do with it. The CFO or managing partner approves a number. The team goes shopping within that number. The goals get defined around whatever the budget can accommodate.

This feels practical. It’s actually backwards.

When budget drives goals, you end up with goals that fit your spending comfort zone — not goals that reflect what your firm actually needs to accomplish. You optimize for what you can afford rather than what you’re trying to achieve. And because the goals were never really defined independently, it’s nearly impossible to evaluate whether the investment worked.

The better approach is to reverse the sequence. Define what success looks like first — specifically, measurably, in terms that matter to your firm. Then build the strategy that gets you there. Then determine what that strategy costs. The budget becomes a function of the goal, not the other way around.

This doesn’t mean budget is irrelevant — of course it matters, and constraints are real. But there’s a meaningful difference between “we can spend X, what can we do with it?” and “we want to achieve Y, what will it cost and how do we phase it?” The second question leads to better decisions, better prioritization, and better outcomes.

It also, not coincidentally, connects directly back to the technical roadmap introduced in Article 2. A roadmap built around clear goals tells you not just what to implement, but in what order, at what cost, and by what milestone. Budget conversations become much easier when they’re anchored to a specific, defensible destination.

Redefining What Success Means for Your Firm

Before you can let goals drive your budget, you have to define the goals. And for most accounting firms, that’s harder than it sounds — because “success” in AI isn’t always obvious.

It’s tempting to borrow success metrics from the industry reports. Firms using AI report 30% faster month-end close. Tax preparation AI reduces processing time by 50-70% for standard returns. Advisory rates are 40-60% higher than compliance work. These numbers are real and meaningful. But they’re averages across many firms in many contexts. They’re not your firm’s goals.

Your definition of success should be grounded in your specific situation: your size, your client base, your workflows, your competitive position, and your capacity for change. Here are the questions that get you there:

What problem are we actually solving?

Success has to be defined in terms of a specific outcome, not a general aspiration. “Use AI more” is not a goal. “Reduce time spent on data entry by 40% within six months” is a goal. The more specific the problem, the more meaningful the success metric.

What does the firm look like when this is working?

Before you start, describe the future state in concrete terms. What are your team members doing differently? What are clients experiencing differently? What’s on the P&L that wasn’t there before? This exercise forces you to move from abstract ambition to operational reality.

How will we know in 90 days if we’re on track?

Long-term goals need short-term indicators. If you’re aiming for a 30% reduction in month-end close time by year’s end, what should you be seeing at 90 days? Defining leading indicators keeps the initiative from drifting and gives leadership something to evaluate before the full investment is made.

What does failure look like, and what’s our threshold?

This question rarely gets asked. It should be the first one. Knowing in advance what would cause you to stop, pivot, or rethink the approach is not pessimism — it’s risk management. It’s also what distinguishes firms that learn from failed pilots from firms that keep funding them out of sunk-cost inertia.

The Bottom Line

The data is clear: firms with a defined strategy succeed at AI adoption at more than twice the rate of firms without one. The foundation of any real strategy is a clear, honest, specific definition of what success means for your firm — built before the budget is set, before the tools are chosen, and before the first implementation begins.

That sequence feels counter-intuitive in an industry that moves fast and rewards action. But the firms widening the competitive gap right now aren’t the ones who moved first. They’re the ones who moved with intention.

Coming Up in Article 5

You’ve built your vendor strategy. You’ve assessed your readiness. You’ve defined what success looks like.

Now it’s time to actually start. Article 5 covers how to pick your first AI use case — why starting narrow wins, how to evaluate your options, and what the firms that get it right do differently in those first 90 days.

Sources & References

G2. “AI Adoption in 2025: 79% of Businesses Prioritize AI in Software.” G2 Learn Hub, April 2026.

UK Government. “Artificial Intelligence Adoption Research.” January 2026.

BILL / NewtonX. Survey of 207 Accounting Firm Leaders. CPA.com, March 2026.

Writer.com. “Generative AI Adoption in the Enterprise.” Enterprise AI Adoption Report, 2025.

McKinsey & Company. “The State of AI: Global Survey 2025.” McKinsey Quarterly.

MIT Sloan. “State of AI in Business 2025.” Massachusetts Institute of Technology.

High Peak Software. “State of AI 2026: Top Industries Driving AI Adoption.” May 2026.

OECD. AI Adoption Research Across Advanced Economies. 2025–2026.

CPA.com. “AI Adoption in Accounting is Rising, But Trust Lags.” January 2026.