Skip to content

AI Assistants in Accounting

In an era where technology rapidly evolves and integrates into every fabric of business, the recent announcements by giants like QuickBooks, Sage, and Xero about integrating AI assistants into their products (links to promotional videos at the bottom) were meant to herald a new dawn for small businesses. At first glance, this innovation seems like a major win, promising to make information access and routine tasks more manageable for small business owners. Yet, beneath the surface of these advancements lies a significant challenge that could potentially undermine the entire endeavour: the quality of data within these accounting systems.

The Core Issue: Data Quality

The bedrock of any AI’s effectiveness is the quality of data it processes. Unfortunately, the reality for many small businesses using these platforms is a landscape littered with incomplete, inaccurate, and inconsistent data. How, then, can we expect AI assistants to guide small business owners effectively?

Imagine an AI assistant tasked with providing financial insights or forecasting based on flawed data. The outcome is akin to building a house on a shaky foundation—it might look good from a distance but is prone to collapse upon closer inspection.

Who do you think will be responsible for underpinning those foundations when they are found to be substandard? You got it – you!

The chatter is loud and clear: clients are consistently left wanting when it comes to the promptness and quality of service from their accountants and bookkeepers. It’s no industry secret that we’re grappling with a bit of a PR nightmare – a mass exodus of professionals and a dwindling interest from the next generation. From this vantage point, the allure of deploying AI Assistants to bridge the widening chasm seems almost logical. However, let’s not confuse a possible solution with the right one. The capability to do something doesn’t inherently justify its execution. Sometimes, the path less travelled by technology might just be the one that leads us back to our true north.

The Capacity Problem

The quality and reliability of client-generated data leaves much to be desired, leading to significant bottlenecks at critical reporting times. This, in turn, hampers our ability to deliver timely and precise advice.

Despite accountants having unprecedented access to client data, the essential information that could facilitate proactive advice and prompt action is frequently drowned out amidst the cacophony of less relevant data.

The absence of standardized workflows further complicates matters, preventing the adoption of best practices. As a result, only the most adept professionals manage to deliver efficient and consistent outcomes, placing an unsustainable burden on our top talent.

Our brightest stars are facing burnout, overwhelmed by their workload. The industry’s inability to adapt and introduce more effective working methods is driving away its most valuable assets.

Accountancy firms are already stretched thin, with capacity being a pivotal issue. The introduction of new digital standards only exacerbates the problem, demanding skills and resources that are in short supply. A staggering 75% of firms report being understaffed.

Moreover, the ACCA, the largest global accounting professional body, has witnessed an 8% drop in student enrolments recently, signalling a worrying trend in the profession’s attractiveness and sustainability. Is it any wonder with 75% of accounting graduates fearing jobs will be taken by robots.

Are AI Assistants a self-fulfilling prophecy?

The Rush for Innovation vs. The Need for Accuracy

The drive to be the first in the market with new technology is understandable in our competitive world. However, when it comes to financial data and the tools businesses rely on for decision-making, accuracy cannot be sacrificed at the altar of innovation. The introduction of AI assistants without a robust mechanism to ensure the data quality they depend on is tantamount to putting the cart before the horse.

The Xero JAX webpage carries a disclaimer stating that the video trailer merely showcases the company’s vision for JAX, hinting that actual products may deviate from this preview. Delving deeper reveals that the beta version is slated for a vague “later” in 2024. Thus, this recent surge in marketing appears to be little more than an idea and an AI-generated glimpse into a potential future.

One must wonder, what’s the real benefit here for the end user and their advisor? It seems to spark more questions than it provides answers, positioning Xero in a light where it’s seemingly just trying to keep pace with competitors rather than offering concrete value.

There is nothing innovative about this. It’s not even real.

Tom Goodwin author of Digital Darwinism wrote this on LinkedIn just a day or two before these announcements and it felt apt. If you haven’t read Digital Darwinism I highly recommend it. It feels like Tom has looked inside my head and made some form of sense of it.

What’s fuelling this frenzied scramble for gold? A potent mix of investor greed and a pervasive Fear of Missing Out (FOMO) among directors and senior executives, all under the intense pressure to meet investor expectations. But one must wonder, has there been any real pause to consider the broader implications of these decisions on the people who operate under these decision-makers? How do the immediate financial gains for a select few weigh against the lasting consequences for humanity as a whole?

The Misplaced Focus: Automating the Problem

There’s an adage in computer science: “Garbage in, garbage out.” This is precisely the risk we run by focusing on automation without first addressing the underlying data quality issues. Automating processes based on erroneous data doesn’t solve the problem; it exacerbates it. This not only misleads businesses but also amplifies the risk of making decisions based on flawed insights.

We have created more data in the last 2 years than in the history of mankind. Key data leading to proactive advice and timely action is often lost in the noise.

Tools like Xenon Connect, Dext Precision, and Xbert represent valiant efforts to address these data quality issues, targeting the accounting market with solutions designed to enhance accuracy. However, their focus on professionals in the field does little to shield end users, the small business owners, from the pitfalls of navigating these complex systems alone and before they converse with their advisor. They will see this as a fast-track route to the answer they need now in a world of immediacy at a perceived lower cost point.

The mere fact that there are tens of products in this filed demonstrates that there is a need to tackle the data quality issues within our accounting ledgers.

Accountability and the Fallout of Errors

One of the thorniest issues in integrating AI into accounting systems is accountability. When errors or omissions occur, the AI chatbot isn’t the one that must answer for the consequences. The burden falls back on accountants and bookkeepers, who must untangle the mess. This not only increases their workload but also strains the trust between them and their clients. In the end, nobody wins.

Accountants find themselves drowning in a sea of client expectations, adrift in an age where the demand for immediate results spares no profession, not even accountancy. Yet, in this race against the clock, we must hold firm to the principle that speed should never eclipse the fundamental need for accuracy. It’s crucial to distinguish between striving for accuracy and chasing perfection; they are not synonymous.

For a more detailed understanding of the subject, please checkout AI in Accounting: Who Carries the can? from Tom Herbert of AcccountingWeb. This quote taken from IBM in 1979 encapsulates the essence of the article wonderfully:

“A computer can never be held accountable.

Therefore, a computer must never make a management decision.”

Moving Forward: A Call for a Balanced Approach

The path forward should be one of balance and caution. While integrating AI into accounting products is undeniably appealing, it’s crucial that we first lay a solid foundation by addressing the quality of the data these systems rely on. Investments in improving data integrity, user education, and error detection should precede or at least accompany the development of AI functionalities.

The aim should be to create a symbiotic relationship between AI and accounting professionals, where each enhances the other’s effectiveness. By ensuring that AI assistants are built on a foundation of reliable data, we can unlock their true potential to transform the world of small business accounting.

Conclusion

While the race towards innovation in accounting technology is commendable, it’s vital that we don’t lose sight of the basics. Data quality is not just an operational issue—it’s the linchpin of trustworthy, effective AI assistance. Let’s not be seduced by the allure of automation before ensuring that the data it relies on is as accurate and reliable as the professionals who are meant to benefit from it.

It’s hardly a secret among us that the bane of our professional existence has been subpar quality. So passionate am I about addressing this Achilles’ heel, that I took it upon myself to create Hindsight—a tool designed to tackle this very plague within the realm of accounting software. This isn’t just a passing interest for me; it’s a mission. Now, when it comes to the buzz around AI Assistants, let’s just say I’m not boarding that train just yet. My conviction lies elsewhere: in the realm of superior accounting products, enhanced training programs, robust standards, and, fundamentally, fostering a workplace that doesn’t just demand our presence but delights in it. That, my friends, is the fertile ground in which I believe our profession will not only grow but flourish.

Additional Resources:

Sage CoPilot Promo Video

Intuit Assist

Xero JAX Promo Video