The Future of Accounting Is Conversational (And Accountants Should Be Excited)

Accounting firms across the United States process roughly 80% of their billable hours on work that requires no professional judgment: data entry, transaction categorization, bank reconciliation, and report formatting. A 2024 Thomson Reuters survey found that accountants spend an average of 4.2 hours per day on tasks they describe as "mechanical." These are professionals who passed the CPA exam, mastered tax law, and built client relationships over decades. And most of their day looks like copying numbers from one system to another.

The fear in the profession right now is that AI will replace accountants. That fear is pointed in the wrong direction.

What actually changes

Consider what AI is already good at: reading bank feeds, matching transactions to categories, reconciling accounts, and generating formatted reports. These are pattern-matching tasks performed against structured data. They are the definition of automatable work.

Now consider what AI cannot do: advise a business owner on whether to take a distribution or a salary, interpret how a new state regulation affects a multi-entity structure, decide whether an aggressive tax position is worth the risk, or look a client in the eye and say "your business partner is stealing from you."

The professional value of an accountant has never been data processing. It has always been judgment. But the profession's economics have been built around selling hours spent on data processing, because that work was unavoidable. When the unavoidable becomes automated, the economics shift. The question is whether that shift threatens accountants or frees them.

I have spent the past three years talking to accountants about this question. The ones who feel threatened tend to define their job by the tasks they perform. The ones who feel excited tend to define their job by the outcomes they create for clients. Both groups are right about what they see. They are just looking at different things.

A different interface

The shift becomes concrete when you change the interface. Traditional accounting software presents spreadsheets, ledgers, and dashboards. The accountant's job is to navigate these interfaces, find the relevant numbers, and synthesize them into meaning. The software organizes data; the accountant interprets it.

Conversational accounting inverts this. Instead of opening a ledger and scanning rows, an accountant asks: "Show me all uncategorized transactions from February." Instead of building a custom report, they say: "Flag any entries that might trigger an audit risk." Instead of manually cross-referencing accounts, they ask: "Which vendors had payment terms change this quarter?"

This is not a cosmetic change. It restructures how accountants spend their time. When gathering data takes seconds instead of hours, the accountant's day fills with the work they were trained to do: analysis, strategy, and client communication.

Think about what this means for how an accountant thinks. With a traditional interface, the accountant has to know which report to pull, which filters to set, which date ranges matter. The interface demands technical fluency in the software itself. With a conversational interface, the accountant just needs to know the right question. The expertise shifts from "how do I get this data out of the system" to "what does this data mean for my client." That is a meaningful difference.

The Tuesday morning test

Here is a practical way to measure the difference. Take a mid-size accounting firm with 200 small business clients. Every month, a staff accountant spends roughly three hours per client on bookkeeping review: categorizing transactions, reconciling bank statements, and preparing financial statements. That is 600 hours of monthly capacity consumed by data processing.

With conversational tools handling the gathering and organizing, that same review takes 30 to 45 minutes per client. The AI has already categorized transactions, flagged anomalies, and prepared draft statements. The accountant reviews, corrects, and approves.

The firm now has roughly 400 hours of monthly capacity freed up. The question becomes: what do those hours become?

Some firms will simply take on more clients, competing on volume. But the more interesting firms will fill those hours with advisory work. A staff accountant who used to spend Tuesday morning categorizing expenses for a restaurant client now spends that time calling the client to say: "Your food costs jumped 12% last month. Two of your suppliers raised prices. Here are three options."

That conversation is worth more to the client than a clean set of books. And it is worth more to the firm than another hour of data entry billed at $85.

The trust problem, solved differently

Accountants have a reasonable objection to this vision: "I review every transaction because that is how I catch problems. If AI does the initial categorization, how do I trust it?"

This objection deserves a serious answer, not a dismissive one. Accountants are professionally liable for the accuracy of financial statements. Their licenses are on the line. Telling them to "just trust the AI" is not a strategy.

The better approach is exception-based review. The AI processes all transactions and presents the accountant with a short list: here are the 15 transactions it was uncertain about, here are the 3 that deviate from historical patterns, here is the vendor it has never seen before. The accountant's review becomes targeted rather than exhaustive.

This is actually a more rigorous review process than the one it replaces. When a human scans 500 transactions line by line, fatigue and pattern blindness set in around transaction 200. The eye glazes. The brain starts auto-approving. When an AI flags 15 anomalies and a human examines each one with full attention, the error-catch rate goes up. The accountant is not doing less quality control. They are doing more effective quality control on a smaller, higher-signal set of items.

There is something counterintuitive here that matters. Many accountants assume that reviewing fewer transactions means doing a worse job. But the opposite is true. A targeted review of flagged items with full context and attention outperforms an exhaustive scan done under time pressure. The quality of attention matters more than the quantity of transactions touched.

The parallel to medicine is worth noting. Radiologists do not examine every pixel of an MRI scan with equal attention. AI highlights areas of concern, and the radiologist focuses their expertise on those areas. No one argues that this makes radiology less rigorous. It makes it more precise.

What the client relationship becomes

The downstream effect on client relationships is significant. Today, most small business owners interact with their accountant in a narrow, transactional way. They send documents, wait for reports, and have a conversation once a quarter about how things look. The relationship is reactive and centered on compliance.

When the accountant is freed from data processing, the relationship can become proactive. The accountant notices a trend before the client does. They call about cash flow concerns before they become cash flow crises. They suggest entity structure changes based on growth patterns they can see because they are not buried in categorization work.

There is a compounding effect here. A proactive accountant generates more value for the client, which makes the client more willing to engage, which gives the accountant more context, which makes their advice better. The relationship builds on itself in a way that transactional compliance work never does.

The conversation between business and accountant shifts from "here are your numbers" to "here is what your numbers mean." That shift changes the accountant from a cost center in the client's mind to a strategic partner. And strategic partners do not get shopped on price the way compliance services do.

The profession's real risk

The real threat to accountants is not AI. It is refusing to let AI handle the work that should not require a trained professional in the first place. Firms that cling to the billable-hour model for data processing will find themselves competing with software that does it faster and cheaper. Firms that redirect their expertise toward advisory will find themselves more valuable than ever.

The economics support this clearly. Advisory services command higher margins than compliance work. Client retention rates for firms offering advisory are roughly double those of compliance-only firms, according to data from the AICPA. And the labor market is pushing in the same direction: the profession is struggling to attract new graduates, and the number one complaint from young accountants is that the work is tedious. Remove the tedium, and the profession becomes more attractive to the people it needs most.

This is not a prediction about some distant future. The tools exist now. The firms that are already making this shift report higher client retention, better margins on advisory services, and, perhaps most telling, happier staff. Junior accountants did not go through five years of education to categorize expenses. They went into the profession to help people make better financial decisions.

At Well, we build the infrastructure that makes this transition practical: AI that handles the gathering and organizing, surfaces anomalies and suggestions, and puts the accountant in the role of reviewer and advisor. The goal is not to replace the accountant. It is to give the accountant back the job they signed up for.

— Maxime