AI-Washing: Why Most "AI" Software Isn't What You Think

AI-Washing: Why Most “AI” in Accounting Software Isn’t What You Think

If you opened any accounting software vendor’s website today and searched for the word “AI,” you would find it everywhere.

AI-powered insights. AI-driven reconciliation. AI-assisted coding. AI-enhanced workflows. Smart features. Intelligent automation. Machine learning at the core.

The language is consistent, the claims are compelling, and the marketing is, it has to be said, very good indeed.

The technology underneath it, in many cases, is considerably less impressive.

This is not a minor distinction. When accounting firms make investment decisions based on AI capability, when they choose platforms, justify costs to partners, and tell clients they are operating at the frontier of the profession, the gap between what is claimed and what is delivered matters enormously. It matters commercially. It matters professionally. And it matters strategically, because a firm that believes it has implemented AI when it has implemented something substantially less sophisticated is a firm that has stopped looking for the real thing.

This article is about how to see the difference. Not as a technology specialist, I am aware that most accountants are not and have no desire to be, but as a professional who needs to make informed decisions about the tools that will define their practice for the next decade.

A Different Vantage Point

Before I go further, it is worth being explicit about where this perspective comes from, because the vantage point matters.

Most commentary on AI in accounting is written by people who came from accounting. They understand the profession deeply, and they understand AI as the profession is experiencing it: as a new, rapidly evolving feature set layered onto familiar platforms.

I came from the other direction.

I have spent over a decade designing and deploying AI and automation systems in some of the most demanding, regulated, and operationally complex environments that exist, financial services, healthcare, critical infrastructure. Environments where AI is not a product feature. It is load-bearing. Where the consequences of a poorly designed system are not a frustrated user or a miscategorised transaction, they are regulatory sanctions, failed audits, and serious operational risk.

I brought that expertise into accounting because the opportunity was obvious and the gap was real. And that origin, AI-first, entering accounting, is fundamentally different from accounting-first, adding AI. The ceiling of what you can build is shaped, more than anything else, by the depth of your understanding of the technology you are building with.

This distinction matters when evaluating any vendor’s AI claims, including ours. The question to ask is not just what the product does, but how deeply the team behind it understands what they have built, and what they could build next.

Start With the Goal. Not the Tool.

Here is the point that I find most consistently missing from conversations about AI in accounting, and the one I want to establish clearly before anything else.

The single most important thing any accounting firm needs to develop is not an understanding of AI. It is an understanding of what they are trying to achieve.

What outcome matters to your firm? What does success look like in two years, specifically, measurably, in terms that your partners would agree on? What is genuinely preventing you from getting there today?

These are not technology questions. They are business questions. And they need to be answered before any technology conversation begins.

AI is one of many possible enablers. In some cases, automating high-volume, repetitive, well-defined workflows, it is an exceptionally powerful one. In others, it is not the right tool at all. Sometimes the problem is a process that needs redesigning before it can be automated. Sometimes it is a data quality issue that no amount of AI can compensate for. Sometimes the friction in the workflow is a people and training issue, or a client communication issue, or a pricing model issue, none of which AI solves.

The firms that are getting the most from AI are not the ones that started by asking “how do we use AI?” They are the ones that started by asking “what are we trying to do?”, and then evaluated AI honestly against other possible approaches to getting there.

AI, although remarkable in what it can do, is not always necessary. It is not always the best solution. And a firm that reaches for AI before it has clearly defined its goal is likely to end up with an impressive-sounding tool that is solving the wrong problem.

Start with the purpose. Let the technology follow from that.

What AI-Washing Actually Means

AI-washing is the practice of applying the language of artificial intelligence to products and features that do not, in any meaningful sense, involve it.

It is not unique to accounting. It has happened in financial services, in healthcare, in retail, in every sector where AI became commercially attractive before the market developed the sophistication to interrogate the claims. In each of those sectors, the same pattern played out: a period of widespread adoption of superficially impressive tools, followed by a reckoning when the results didn’t materialise, followed finally by the emergence of tools that actually worked, built by companies that understood what AI genuinely required.

Accounting is in the middle of that pattern right now.

The tools being marketed as AI in the accounting software market exist on a very wide spectrum. At one end is genuine artificial intelligence, systems that learn, adapt, reason, and act across complex multi-step workflows with meaningful autonomy. At the other end is a predefined rule that fires when a transaction matches a keyword and puts it in a category.

Both of these things are currently being sold to accounting firms as “AI.”

The Maturity Spectrum - Where Things Actually Sit

To understand what you are actually buying, it helps to understand AI not as a binary, present or absent, but as a maturity spectrum with meaningfully different levels.

Level 1, Rule-based automation Predefined logic. If this transaction contains the word “HMRC,” put it in the tax category. If the amount matches last month’s standing order, reconcile it. No learning involved. No intelligence involved. Just rules, applied consistently at speed. This is useful. It is not AI.

Level 2, Pattern recognition. The system learns from historical data. It observes that you previously categorised transactions from a particular merchant in a particular way, and suggests the same categorisation next time. More sophisticated than Level 1, but still fundamentally pattern matching trained on your own historical behaviour.

Level 3, Predictive analytics. The system uses historical patterns to forecast future states. Cash flow predictions. Revenue forecasts. Anomaly detection based on statistical deviation from expected ranges. Genuinely useful analytical capability, but still extrapolating from the past rather than reasoning about the present.

Level 4, Natural language interaction. The system can understand and respond to queries in plain English. You ask a question; it retrieves relevant data and presents an answer conversationally. This is where the large language model features in most accounting platforms currently sit. Useful for certain queries. But still dependent on data quality and the precision of the question being asked.

Level 5, Autonomous decision-making. The system makes decisions independently, within defined parameters, and acts on them without prompting. This is where meaningful workflow transformation begins. Very few accounting tools are genuinely operating at this level.

Level 6, Agentic AI. The system can plan across multiple steps, use external tools and systems, handle situations it has not encountered before, and pursue a defined goal with genuine self-direction. This is real agency. It is coming to accounting. It is not yet widely present in the market in production-ready form, whatever the marketing materials say.

Most embedded “AI” features in mainstream platforms sit firmly at Levels 1 and 2. Some of the more advanced tools are reaching Levels 3 and 4. Levels 5 and 6 remain largely the territory of specialist, purpose-built platforms. The enterprise world I came from operates at Levels 4, 5, and 6 as a matter of routine. Accounting is not behind by months. It is behind by years.

How to Spot It in Practice

The good news is that AI-washing is not difficult to identify once you know what to look for. The following questions will separate genuine capability from marketing language in any vendor conversation.

“Can you show me a workflow that runs end-to-end without a human initiating each step?” A genuinely automated workflow receives a trigger, processes data, makes decisions, and produces an output without being prompted at every stage. If the demo shows a series of smart suggestions that a human then acts on, that is Level 2 automation, not a workflow.

“How does the system handle something it hasn’t seen before?” Rule-based systems fail badly on unfamiliar inputs. A genuinely intelligent system can handle novelty, a transaction from a new vendor, an unusual combination of conditions, an edge case outside the training data. Ask for a live demonstration with an unusual input and watch what happens.

“What is the system actually learning, and how do we know?” Many systems claim to “learn from your corrections.” What this usually means is that the correction is stored and applied next time the same pattern appears. That is memory, not learning. Ask the vendor to explain specifically what the model is learning, how frequently it is updated, and how they measure improvement. Vague answers are not answers.

“Is this feature built on your own model or a third-party API?” A significant proportion of the “AI” features in accounting software are third-party large language model APIs, typically OpenAI or similar, with a prompt and interface built on top. The accounting company’s competitive advantage rests on the quality of the interface, not any underlying AI development. Worth knowing.

“If I correct the same error next month, will the system have learned from it?” If the answer is no, if the same correction needs to be made repeatedly, that is automation with a marketing label, not a learning system.

The Bolt-On vs. Built-For Distinction

There is a structural dimension to this that goes beyond individual features.

The accounting software market currently divides, at the architecture level, into two fundamentally different categories. Traditional platforms built for accounting that have added AI capabilities as the market has demanded them. And platforms built specifically for AI-native workflows from the ground up.

This distinction determines the ceiling.

A platform built on a legacy architecture, even one with significant AI investment bolted on top, faces structural constraints. The data model, the processing pipeline, the integration layer, the workflow logic: all were designed for a world where humans initiate and complete every significant action. Adding AI to that architecture is impressive engineering. But it is engineering in service of an existing paradigm, not a new one.

A platform built AI-first, where the architecture assumes that systems will reason, decide, and act autonomously as the default, is a different thing entirely. Its ceiling is categorically higher.

This is not unique to accounting. In every industry where I have deployed AI at scale, the firms that achieved genuine transformation did not do it by adding AI to what they already had. They did it by asking what the process should look like if AI were the starting assumption, and then building toward that.

AI-Washing Has a Short Shelf Life, But What Replaces It Matters More

Here is a prediction worth making, because it changes what accounting firms should be optimising for right now.

AI-washing, as a problem, will largely resolve itself within the next year or two. Not because vendors will suddenly become more precise in their language, but because AI will become so commoditised, so embedded in every product across every category, that the label will cease to carry any meaningful signal at all. When every piece of accounting software natively includes some element of AI as standard, claiming “AI” will be about as differentiating as claiming “cloud” is today.

What will matter then, what the firms building for the future should be thinking about now, is not whether the tools have AI, but whether the AI has been shaped to work for them, specifically.

Personalisation. Localisation. The ability to configure AI to reflect your firm’s way of working, your client relationship model, your terminology, your risk appetite, your pricing approach. AI that has been trained, or fine-tuned, or prompted, to sound like your firm, to understand your clients’ businesses, to apply your professional judgment at scale rather than a generic model’s best approximation of it.

This is the missed opportunity that most accounting firms are not yet thinking about. They are asking whether the platform has AI. They should be asking whether the AI can become theirs.

The firms that will lead this market in five years will not simply be the ones that adopted AI earliest. They will be the ones that made it their own, that built proprietary data advantages, personalised their client-facing outputs, and created an AI-enabled experience that a client or a staff member cannot find anywhere else.

That is the real differentiator. And it is available to any firm that thinks clearly about it now, before the commodity wave arrives and homogenises the market.

A Word on Fairness

I want to be clear about something, because intellectual honesty matters in a conversation like this.

The platforms currently operating at Levels 1 and 2, Xero, QuickBooks, Sage and others, are not being dishonest when they describe their features as AI. In many cases, the features genuinely involve machine learning, even if it is basic. The language of AI has expanded to encompass a wide range of capabilities, and the vendors are operating within the conventions of that expanded definition.

What they are doing, intentionally or not, is allowing that broad definition to obscure meaningful differences in capability. When every product claims AI, the term ceases to be informative. The sophisticated buyer has to look past the label to the substance, and most accounting firms, pressed for time, trusting in established brands, reassured by familiar interfaces, do not do that.

The point is not that established platforms have nothing to offer. They do. The point is that choosing a platform because it says “AI” on the box, without understanding what level of AI capability that represents and whether it matches the problem you are trying to solve, is a decision made without adequate information.

And in an environment where that decision will shape your firm’s competitive position for the next decade, adequate information matters.

What to Do With This

The practical takeaway from this article is not “distrust AI claims.” It is “interrogate them”, and before you interrogate the vendor’s claims, interrogate your own assumptions.

What are you actually trying to achieve? Is AI genuinely the right tool for that goal, or are there simpler, faster, or more effective approaches? If AI is the right tool, what level of AI capability does the problem actually require? And is the platform you are evaluating genuinely operating at that level, or is it operating at Level 2 and telling you it is Level 5?

These are not difficult questions. They are just questions that the market does not currently encourage you to ask.

Start asking them. The difference between the answers, between what is being sold and what is being delivered, is the difference between efficiency and transformation. Between doing what you have always done slightly faster, and doing something your competitors genuinely cannot.

That distinction is worth making. And right now, the market is not making it clearly enough.

| Daniel Lawrence is the Founder of Bots For That and creator of the Beanieverse platform, a suite of AI-native powered tools for the accounting and bookkeeping sector. With over a decade of experience deploying enterprise automation and AI in financial services, and other highly regulated industries, he writes about AI transformation in accounting from the outside in. |