AI reality

The Ai Gap: Promises and Reality in Accounting

Ai gap? What Ai gap?

If you listen to software vendors, conference speakers, and industry analysts, Ai in accounting is already transforming the profession. The tools are ready. The results are proven. All you need to do is adopt the technology and watch your efficiency soar.

If you talk to the people actually doing the work, you hear a very different story. The Ai accounting features don’t work as advertised. The time savings never materialize. The technology creates new problems while solving old ones. What looked simple in the demo turns complicated in real life.

This gap between what executives and vendors say about Ai and accounting and what employees and practitioners experience is one of the least talked about facts of Ai implementations. It’s not just that the technology falls short of expectations. It’s that the expectations themselves are based on a version of reality that doesn’t exist in most businesses yet.

Executive Messaging Doesn’t Match AI Implementation

Company leaders and software executives often champion Ai as a huge efficiency booster. They talk about hours saved, processes streamlined, and competitive advantages gained. Their enthusiasm is genuine, but it’s also shaped by their distance from how the work is actually done.

The numbers reveal a massive disconnect. Roughly 40% of company executives claim Ai saves them more than eight hours of work per week. But when you talk to their employees, two-thirds of non-management staff say they save less than two hours or no time at all. That’s not a small difference in perspective. That’s two groups of people working in the same business having completely opposite experiences with the same technology.

The Ai gap gets wider when you look at how Ai affects workload. More than 70% of corporate executives report feeling excited about Ai. At the same time, roughly 70% of rank-and-file employees feel anxious or overwhelmed by it. And for good reason. About 77% of employees say Ai has actually increased their workload rather than reduced it. Instead of eliminating work, Ai has created what some call an “Ai tax” or “Ai rework,” a negative impact where employees spend significant time correcting Ai errors to meet employer standards. Workday estimates that 37% of the savings created by Ai goes right back in to correcting, reworking, and amending Ai output.

To quote Andrew Kershaw, Workday’s Group General Manager for the CFO, “AI only delivers real value when people know how to use it well.”

This isn’t theoretical. At The Numbers Advisors, we recently took on a client who had been with another accounting firm for a couple of years. The Ai-powered tools that firm used created problems the bookkeeper didn’t fully understand. Rather than admit they lacked the expertise to fix it, the bookkeeper applied band-aid solutions. This resulted in a series of mismatched transactions and adjusting journal entries that compounded over time. We’re now working through a significant cleanup project, the Ai tax, to untangle what the Ai and the bookkeeper created together.

In another case, we took on a client who had tried to implement Ai bookkeeping and the experience went so badly that they stopped using QuickBooks altogether and went back to manual spreadsheets. The promise was efficiency. The reality was so frustrating that they chose to revert back to a slower, more labor-intensive method because at least they could trust the results.

A business owner we know received a retirement notice from their Controller and decided they could replace that position with a bookkeeper at half the salary and “just use Ai to do the rest of the work.” The reality turned out far different from that plan. In the end, the business owner had to hire another Controller at the same rate and implemented very little Ai. The business didn’t fully understand what Ai was capable of, and was simply parroting what they’d heard others say.

What is “Ai washing”

“Ai washing” is when companies either overstate the benefits their business has gained as a result of implementing Ai, or claim that the negative performance in their business is a direct result of Ai being implemented by their clients or competitors. In both cases, Ai is blamed without the company providing any real detail or proof.

An example of “Ai washing” would be Block’s layoffs in the Spring of ’26. In the footnotes of their announcement, they admitted to overhiring.

Why does Ai accounting not work yet?

First, you need someone who fully understands the workflows of every process that affects accounting and who also understands how each piece of software actually works in order to successfully implement Ai in accounting. Those combined talents are rare right now.

Second, implementing Ai isn’t a light switch you flip. It’s a phased approach where the result of each step has to be verified and validated before it can be trusted.

Third, no two businesses are alike, and this reality hasn’t been fully accounted for in the Ai. Remember, Ai isn’t very good with creativity yet. It follows patterns. When your business doesn’t fit the pattern, the AI struggles.

Finally, the business owner didn’t understand all of this, so they had no idea what they were actually asking for when they said “just use Ai.”

The Ai Gap: Executive Vs. Employee Experience

The disconnect between leadership and staff experience with Ai creates real problems for implementation. It’s not just about different perspectives. It’s about decisions being made based on one reality while work gets done in another.

While roughly 78% of businesses reported using Ai in some capacity by 2024, only 1% of leaders describe their organizations as “mature” in Ai deployment. Most companies are stuck in pilot phases or isolated experiments. In late 2025, about 30% of Ai initiatives were expected to be abandoned after the pilot stage because of poor data quality, unclear business value, or cost overruns. Think of the costs this sort of Ai gap puts on a business when they have to backpedal. We’re still waiting on the public announcements…

Despite heavy investment, 56% of companies report they’re getting nothing out of Ai yet. A similar percentage of CEOs say they have not yet seen real financial returns. But the pressure to adopt remains intense. About 64% of CEOs admit they are investing in new Ai technologies before fully understanding their value because they fear falling behind competitors.

If these are the results CEOs are seeing across the entire business landscape, why would the results be much different when we look at using Ai in accounting?

This creates a strange situation where companies are publicly promoting how well they use Ai while internally struggling to make it work. Some experts suggest that certain firms are “Ai washing” their business decisions, using Ai as a convenient story for other corporate actions. For instance, some companies may blame job cuts on Ai to signal cutting-edge adoption to investors when the real reasons might be overhiring or missing financial targets.

In accounting and bookkeeping specifically, “Ai washing” shows up in other ways too. It’s become an easy explanation when things go wrong. A bookkeeper who doesn’t fully understand what happened can blame Ai for errors rather than admitting they lack expertise in either the technology or the accounting process. Unfortunately, many business owners don’t have the time to sit with their accountant or bookkeeper and be walked through what happened and how the process broke down. At The Numbers Advisors, we make a point of inviting clients to see how we discovered a problem so they can understand it at a fundamental level. That transparency matters.

There’s also a strategy gap between what leadership thinks exists and what employees actually have to work with. About 81% of executives believe their company has a clear Ai policy. Only 28% of individual contributors agree. Furthermore, 75% of frontline workers report receiving insufficient Ai guidance from leadership. This means the people actually using the tools don’t have clear direction on when to trust Ai output, when to verify it, and when to override it entirely.

The emotional experience differs sharply too. Executives feel excited about possibilities. Employees feel anxious about workload, job security, and being held responsible for Ai mistakes they didn’t make. One group sees opportunity. The other sees risk.

For small business owners, the Ai gap plays out differently but just as significantly. The owner attends a conference where Ai is heavily promoted. They read articles about competitors using Ai. They get sales calls from software vendors making impressive claims. They come back to their business wanting to implement these tools. But the person actually doing the work knows it won’t work with their messy data, inconsistent processes, or the specific quirks of how their business operates.

The pressure to adopt Ai is real for small business owners. Business and industry are moving in that direction. Vendors are pushing it. Competitors are talking about it. But willingness and ability vary widely. Some owners want to dive in themselves. Others prefer to let someone else handle it while having little understanding of what it actually takes to make it happen. Both approaches can lead to problems when expectations don’t match reality.

The fundamental issue is that implementing Ai for accounting, or any other part of a business, requires expertise, time, careful planning, and realistic expectations. But the public messaging around Ai suggests it’s fast, easy, and delivers immediate results. That messaging shapes how business owners and executives think about Ai. And when their experience doesn’t match those expectations, they often blame themselves, their staff, or their data rather than recognizing that the gap between promise and reality is industry-wide.

Sources and References: