Earmark CPE

Earn CPE Anytime, Anywhere

Digits

Your AI Isn’t the Problem—Your Ledger Is

· ·

You export a report from your accounting software, spend 20 minutes cleaning up the data in a spreadsheet, paste it into ChatGPT, and ask about vendor spending trends. The AI immediately asks for more context that wasn’t in your export. Back to the source system, pull another report, reformat, and upload again. The cycle repeats.

If this loop feels painfully familiar, you’re not alone.

In a recent Earmark webinar, Megan Reid, Product Specialist at Digits, demonstrated exactly why this workflow keeps breaking down and introduced a technology designed to eliminate it entirely. The webinar unpacks what’s really limiting AI’s usefulness in accounting and shows a fundamentally different approach.

The core message is that Model Context Protocol (MCP) eliminates the export-reformat-upload cycle, but only if your ledger is truly AI-ready. As Megan explained, your financial data architecture now determines whether AI delivers reliable insights or confidently wrong answers.

 

The traditional workflow is fundamentally broken

Let’s walk through what Megan calls the “bolt-on” AI workflow. Most accountants know this workflow far too well.

  1. You export a report. The moment you hit that button, your data freezes in time. You’re working with a snapshot, not live information.
  2. Manual cleanup begins. You’re renaming columns, adjusting formulas and reformatting data. Every manual touch introduces potential errors.
  3. You send the cleaned data to your AI tool. It analyzes what you’ve given it and generates an answer, but it can only see that specific export. No additional context or visibility beyond that snapshot.
  4. The AI needs more information. Maybe vendor history or prior period details that weren’t in your original report. You can’t answer without going back to start the entire process over.

“This isn’t a limitation on the AI,” Megan emphasized during the session. “It’s a limitation on the data pipeline feeding into the AI.”

Megan introduced a powerful concept when she explained that AI acts as a megaphone for your data. When the input signal is clean (i.e., real-time, well-structured, and consistently categorized), AI generates accurate analysis and trustworthy insights. When the signal is poor (i.e, outdated data, inconsistent vendor names, and incomplete transactions), the AI still produces an answer.

“A wrong answer delivered with confidence,” Megan noted, “is often worse than no answer at all.”

The accounting profession recognizes this shift. The Journal of Accountancy states that “the profession must pivot from doing to supervising when AI does the work.” As one industry publication put it, “The industry is shifting from manual data entry to automation, where the accountant’s job is less about performing repetitive tasks and more about defining the logic once and letting the system run it.”

But you can’t supervise what you don’t understand. And you can’t get reliable outputs from a broken data pipeline.

Enter Model Context Protocol

So what breaks the cycle? That’s where MCP comes in.

At its core, MCP is an open standard letting AI tools like Claude, ChatGPT, and Cursor connect directly to live data sources. Megan offered a perfect analogy: think of MCP as USB-C for AI. Before USB-C, every device needed a different cable. USB-C created one standardized connection. MCP does the same for AI and data.

The AI doesn’t change. It simply gains direct, permission-based access to your financial information through a standardized connection. No exports, uploads, or stale spreadsheets. The AI reads your live ledger in real time with full context.

But Megan stressed MCP is only as powerful as the ledger it connects to. A direct pipeline to messy data just delivers messy answers faster.

How do you know if your ledger is ready? Megan presented three diagnostic questions:

  1. Are transactions sitting in a queue? In traditional systems, bank feeds arrive uncategorized, waiting for manual review and posting. If your books take two weeks to close, your AI operates on two-week-old information. “It’s difficult for business owners to make real-time decisions using old data,” Megan explained.
  2. Does your system know “Uber” and “Uber Technologies Inc.” are the same vendor? To humans, it’s obvious. To traditional ledgers, they’re separate text strings. “This isn’t a data entry problem,” Megan clarified. “It’s a data architecture problem.”
  3. Is the tool itself a bottleneck? When slow page loads, endless clicking, and constant manual saves slow you down, data falls behind.

If you answered yes to any of these, your ledger isn’t AI-ready, regardless of how sophisticated your AI tools are.

The four pillars of an AI-ready ledger

Megan outlined what an AI-native ledger actually looks like through four pillars:

Real-time processing

Transactions are automatically categorized and posted as they arrive, not sitting in a queue. With Digits, for example, transactions are “continuously being posted, reviewed, reconciled, and reflected in your financials in real time.”

Object-oriented data

Every vendor, customer, and category is stored as a structured object. Variations from the same vendor are treated as a single entity, providing AI and accountants with a reliable foundation for analysis.

Autonomous bookkeeping

The system handles routine bookkeeping automatically, allowing accountants to focus on “reviewing, supervising, advising, rather than manually processing those transactions.”

Accessibility

Even the best technology has limited impact if firms face barriers to adoption. As Megan noted, firms shouldn’t have to “navigate complex pricing models, marketplace restrictions, and AI licensing costs just to take advantage of these modern tools.”

“AI readiness isn’t about having access to AI tools,” Megan summarized. “It’s about having a financial system that can provide accurate, structured, current information for those tools to work with.”

Seeing MCP in action

Theory is one thing. Watching it work is another.

During the live demo, Megan showed exactly what MCP-powered accounting looks like using Claude’s desktop app. Setup was surprisingly simple. You open Claude, navigate to connectors, search for Digits, and add it. No complex configuration needed.

With the connection live, Megan demonstrated real-world scenarios using a demo client:

  • Vendor spend analysis. She asked Claude to review which vendors grew most over three months and flag unusual spending. Within minutes, the AI identified that payroll scaled smoothly with headcount, spotted a November bonus spike, and highlighted fast-rising smaller vendors. No exports or reformatting. Just direct questions and detailed answers.
  • Budget creation. She requested a 2026 budget based on two years of historical data. The AI produced a complete budget showing prior actuals, year-over-year changes, and projections with adjustable assumptions, like testing a 40% revenue growth scenario. Everything was interactive and exportable to Excel.
  • Budget-to-actuals reporting. Building on that budget, she asked for Q1 comparisons. The AI generated monthly trends, variance analysis, and actionable recommendations. It didn’t just show that revenue was below budget and expenses were over. It identified specific areas to review, like legal costs and growth assumptions.
  • Expense optimization. When asked to identify potential cuts, the AI flagged a $1,600 charge for an HR tool that likely overlapped with the demo company’s payroll software. It also spotted a redundant AI bookkeeping subscription that duplicated Digits’ capabilities. These insights normally require hours of manual vendor analysis.

All of this happened in minutes through MCP’s direct, secure access to the live ledger.

The possibilities extend further. As Megan explained, you could connect MCP to multiple tools and ask cross-functional questions like, “Look at the last call I had with Megan’s demo client. What were the things they pointed out and compare that to the actual financials over the last three months? Help me identify what I should highlight with my client on our next call.”

The ledger is your constant. Everything else is variable.

“The best AI tool for your accounting is whatever you prefer,” Megan said in closing. “The best ledger for your AI tool is one that agents love most.”

AI tools will evolve and multiply. The ledger’s quality, structure, and accessibility are the constant determining whether any of them deliver value.

The shift isn’t just technological. As Tom Hood, Executive Vice President, Business Growth and Engagement at AICPA, noted, this is “an inflection point where finance leaders agree that people drive transformation success—mindset, skills, and leadership, not technology alone.”

Your role is evolving from data processor to financial supervisor and strategic advisor. Success means understanding the technology well enough to oversee it effectively.

Here’s where to start:

  • Audit your current workflow. How much time do you spend exporting and reformatting before AI can help? That’s your friction baseline.
  • Apply the three diagnostic questions. Check for queued bank feeds, unresolved vendor duplicates, and tool-created bottlenecks.
  • Evaluate against the four pillars. Real-time processing, object-oriented data, autonomous bookkeeping, accessibility. If your ledger can’t check these boxes, your AI outputs will always be limited.
  • Start experimenting. Connect an MCP-enabled tool and run some prompts. Even using a demo environment will shift your understanding of what’s possible.
  • Build your vocabulary and mindset. Understanding how to evaluate and supervise AI workflows is now a core professional skill.

The traditional export-reformat-upload workflow is broken because the data pipeline starves AI of the context it needs. MCP fixes the pipeline, but only if your ledger is truly AI-ready.

Watch the full on-demand webinar to see the live demos, walk through the diagnostic framework, and start building your firm’s AI-readiness roadmap.

Stop Losing Money on Cleanup Work by Automating the Parts That Don’t Need You

· ·

Cleanup and catch-up work is among the most in-demand services accounting firms can sell, and among the hardest to deliver profitably. That was the starting point for a recent webinar led by Megan Reid, a 15-year accounting veteran who started in Big Four, moved through private industry, and now works on the firm enablement team at Digits.

In the webinar, Megan demonstrated how AI-native accounting tools can transform cleanup engagements from time-intensive projects into scalable service offerings. She built a client file from scratch, imported raw PDF bank statements, and walked through an entire cleanup workflow in real time.

Why cleanup work kills profit margins

“Cleanup is obviously valuable work and it’s hard to scale,” Megan said, framing the core challenge clearly. New clients almost always arrive with some sort of mess to clean up. Maybe you have 18 months of uncategorized transactions or transactions that haven’t been posted from the bank feed. You want to take the engagement, but you know it’s going to be hard to make it profitable.

“We always uncover more skeletons in the closet than we think,” Megan noted during the demonstration. If you’re billing fixed fees, you get squeezed by unpredictable hours. Clients want fast turnarounds. Your teams are leaner. “You’re asked to do more with less,” she said.

“Business owners need that work to be done,” Megan pointed out. But the question is “whether or not your workflow lets you take them profitably.”

Breaking down a traditional cleanup shows where the hours go:

  • Gathering data
  • Importing it or connecting feeds
  • Categorizing tons of transactions
  • Reconciling accounts
  • Resolving exceptions
  • Making adjusting entries
  • Reviewing everything with your client
  • Delivering the final report

“In a typical 12-month cleanup or catch-up, you spend the majority of your time categorizing and reconciling transactions,” Megan explained. These tasks are also “the most repetitive, pattern-based parts of the job, which is exactly what AI is good at.”

From blank file to categorized transactions

Megan started her demonstration with a completely blank client file, essentially just an empty ledger. She then showed how to handle a common scenario in which a new client hands over a stack of PDF bank statements with no bank login credentials.

She dragged and dropped the first PDF bank statement directly into Digits. “It is extracting all that data from the bank statement, booking it and categorizing it as well,” Megan explained as the system processed the document.

The AI extracted transactions, identified vendors and customers (called “parties” in Digits), populated company logos and descriptions, attached website links, and categorized each transaction into the appropriate account. Megan noted the system pulls from models trained on “more than 800 trillion dollars’ worth of transactions.”

After uploading statements for June through October, hundreds of transactions flowed in. When processing finished, only 12 were flagged for review. “Instead of manually clearing bank feeds,” Megan said, “come here and look at the exceptions.”

These were transactions that required confirmation. Megan clicked into one from Swift Courier Services. The AI suggested “contractors and consultants.” She confirmed it with one click.

From there, the system natively learned from that categorization. It immediately found two similar transactions and offered to update them together. The exception list dropped from 12 to 8 in seconds.

Bank reconciliation without the manual work

Megan demonstrated three ways to get bank statements into the system for reconciliation. You can connect directly to banks like Mercury, Wells Fargo, Chase, and US Bank, which pull statements automatically via API. You can drag and drop PDF statements anywhere in the product. Or you can use email ingestion, where each client gets a unique email address to forward statements.

She uploaded the June statement by dragging it onto the reconciliation screen. The system read the PDF, extracted every line item, and verified each against the ledger. Megan explained that the system uses “pixel bounding boxes” to match statement entries to ledger entries.

June needed one manual step: adding a beginning balance entry that the system couldn’t infer without a connected bank account. Megan created the entry directly in the reconciliation screen. “Unlike legacy systems, where you may have to have three different tabs open and make changes and then come back and refresh, everything can be done directly in here.”

Then she uploaded July’s statement and navigated away. When she returned, it was done. “The statement was uploaded by me. The auto reconciliation was kicked off by Digits and even finalized by Digits,” she showed in the timeline view.

For larger cleanups, Megan recommended uploading multiple statements at one time. Handle any beginning balances in the first month, then subsequent months often complete automatically.

Review tools that surface what matters

Even with AI handling categorization, accountants still need to review and sign off. “It doesn’t replace the accountant. It just removes that tedious work so that you can focus on those judgment calls,” Megan emphasized.

She demonstrated several review approaches. The general ledger view shows all transactions organized like a trial balance, including assets, liabilities, equity, revenue, and expenses. You can filter by status, amount, source, department, or location. Bulk updates work on hundreds of transactions at once.

Megan said the vendors and customers views are her favorite. They each flag two critical items:

  • New vendors or customers: Any vendor (or customer) the AI sees for the first time in your selected period
  • Split categorizations: Vendors (or customers) whose transactions appear in multiple categories

“I just need to have eyes on things it has not seen before,” Megan explained. Even if the AI categorized with high confidence, you have final review and say on how it was categorized..

For transactions needing client input, the collaboration happens in one place. Megan showed how to comment on any transaction: “Hey client, what is this for?” The client receives an email with a link, can respond directly in Digits or reply to the email, and the response appears on the platform. “All the collaboration is centralized in one location,” she said, “instead of you having to manage a ton of emails and download Excel files.”

Delivering professional reports, not data dumps

The final step Megan demonstrated was creating custom reports. While the financials inside Digits update live as transactions flow in, cleanup engagements need a formal deliverable, a static document that locks the numbers in place.

Megan built one on screen. She added a cover page, used AI to draft an executive summary, embedded links to the client’s checklist, and configured the financial statements with period comparisons and trend lines. The system includes “hover to discover” insights that show period-over-period changes and what drove them.

When you need to make adjustments after sending a draft, you create a new version. “Any adjustments you’ve made in Digits will then update directly to this report,” Megan explained. Publishing the final version removes the draft watermark and notifies the client.

The platform tracks everything, including when you created the report, when you published it, when the client viewed it, and all comments from either party. You have a complete record of the deliverable and the conversation around it.

“We’ve done 12 months of cleanup in an hour and a half instead of days,” Megan concluded.

What this means for your firm

The key takeaways from Megan’s demonstration show how cleanup engagements can become profitable:

  • AI categorizes the vast majority of transactions automatically, flagging only true exceptions
  • Bank reconciliations can run automatically when you upload statements
  • The system learns instantly from every correction without rules to build or maintain
  • Your time shifts to reviewing anomalies, making judgment calls, and delivering polished reports

One practical consideration came up during Q&A. When asked about importing messy QuickBooks Online data, Megan confirmed that direct QBO migration exists but cautioned, “You maybe don’t want the AI to learn off of really messy data. You maybe just want to start fresh.” The system uses imported data for baseline training, so starting clean might make more sense for particularly messy files.

For firms trying to grow, this changes the economics of client acquisition. Every prospect with messy books becomes an opportunity rather than a capacity problem. When you can handle cleanup work profitably, predictably, and consistently, you can say yes to more engagements while maintaining margins.

Watch the full on-demand webinar to see Megan’s complete demonstration from blank file to published financials. If you have cleanup engagements in your pipeline right now, consider what your workflow could look like when the repetitive work is automated.

The Month-End Close Is Accounting’s Biggest Bottleneck. Here’s How AI Is Dismantling It

· ·

The day before a tax deadline, and accountants from Miami to Vancouver, Portland to New York, logged into a CPE-eligible webinar to learn something that could fundamentally change how they work. The webinar showed how these professionals can shrink the most time-consuming part of their month-end close, like reconciliations, transaction coding, and bank statement chasing, from days to minutes.

Megan Reid, product specialist at Digits, led the session, and she brings a unique perspective. She’s an accountant with 15 years in the trenches, starting at a Big Four firm, moving through banking and construction, and now helping firms build what she calls an “AI-native” practice. As she put it to the audience, “As accountants, we want to be able to serve more clients, provide better service, and do so quickly and efficiently.”

The traditional month-end close is accounting’s biggest bottleneck. It’s that manual grind through booking transactions, reconciling statements, updating schedules, reviewing anomalies, and (if there’s time left) analyzing the numbers and creating the reports clients care about. “It’s a manual, tedious, time-consuming process that honestly leaves a lot to be desired for both the business owners and the accountants,” Megan said bluntly. 

But what if you could flip that entire workflow? What if instead of reviewing every transaction, you only touched the ones AI couldn’t confidently handle? That’s exactly what Megan demonstrated live, showing how AI-native platforms transform the close from a compliance chore into an opportunity for real advisory work.

The bottlenecks killing your efficiency

Before diving into solutions, Megan mapped out where the traditional close breaks down. You start in QuickBooks or your ledger of choice, but quickly find yourself bouncing between Excel, browser tabs for vendor research, your close management tool, and who knows what else. “Not only are you managing the work across all these multiple platforms,” she explained, “you’re also spending time validating sync accuracy, troubleshooting issues, and making sure the data moves seamlessly throughout the various systems.”

Each phase has its own special frustrations:

  • Manual data entry and rule management introduce human error
  • Fighting with bank access and chasing clients for statements
  • Disconnected tools for AP, credit cards, and close management
  • Team members use different processes, causing rework and confusion
  • Manual journal entries pile up at period-end

As a result, most of your time goes to necessary but low-value tasks, leaving little room for the analysis and insights your clients actually hired you to provide.

How AI learns your way of doing things

The shift to AI-native platforms involves intelligence that learns and adapts. When Megan pulled up the demo client in Digits, she showed hundreds of transactions the AI automatically categorized. Only eight were flagged for review.

“How does it know how to categorize transactions?” she asked, anticipating the obvious question. The answer lies in three layers of learning.

First, there’s client-level learning. When you correct a categorization for a specific client, the system learns instantly. “If you review something for a brand new client and you say, ‘nope, you categorized this to software, but I actually want it to be cost of revenue,’ Digits learns from that instantly,” Megan explained.

Second, there’s firm-level learning. The system recognizes patterns across your entire client base. If the system does not have the client-level layer of knowledge, it falls to the firm-level. “How has my firm done this across all of my clients? It automatically applies your firm’s unique value to your client base.”

Third, when a transaction is entirely new, proprietary models trained on billions of dollars’ worth of transactions make the call.

During the live demo, Megan reviewed a U.S. Patent and Trademark Office transaction the AI thought might be taxes. She looked at the suggestions (taxes, legal, or a new intangibles account), selected “Legal,” and clicked save. The system immediately found two similar transactions and updated them automatically. The review queue dropped from eight to five in seconds.

But what really eliminates busywork is the AI agents run 24/7 in the background, researching vendors and populating details. “None of this has been populated manually,” Megan showed, clicking through a vendor profile complete with name, logo, description, and related websites. “We’re essentially researching them and populating all of the data for you.”

Bank reconciliation without the chase

If transaction categorization is tedious, reconciliation might be even worse. You know the drill: fighting for bank access, emailing clients for statements, then manually comparing the ledger to the statement line by line.

Megan demonstrated the “happy path” first. Digits pulled a Mercury bank statement via an API, automatically kicked off reconciliation, matched every transaction with pixel-level precision on the PDF, confirmed the ending balance, and finalized everything. Zero human touches required.

“Some firms we work with actually say, ‘I uploaded six months of bank statements and just watched them finalize one by one. And I didn’t do anything,'” Megan shared.

When auto-reconciliation can’t finalize completely, it doesn’t leave you guessing. The system flags specific issues, such as:

  • Missing transactions that exist on the statement but not in the ledger (one click to create)
  • Date mismatches where something cleared May 31 but hit the ledger June 1 (one click to adjust)
  • Unsettled items like checks that haven’t cleared yet

For banks without API access, such as small credit unions, you simply drag and drop a PDF statement. During the demo, Megan dragged a statement into the system and watched it extract data and start reconciling in seconds.

She took it further with a cleanup scenario. Starting with a brand-new bank account, she imported a PDF statement. Within moments, 14 transactions appeared as uncategorized. Seconds later, the AI had populated every vendor name and category without a single manual input.

Turning saved time into client value

Speed alone isn’t the point. As Megan emphasized, “the compliance and the month-end close is really just a means to an end,” the end being insights and value for clients.

The dashboards in Digits default to the current month because, as Megan noted, “knowing something two months late doesn’t usually help.” Every metric is live and drillable. Click into gross income, and you see the definition, calculation, and every underlying transaction. Your clients finally understand how you arrived at the numbers.

Each client gets customized dashboards. “Maybe you have a client that’s like, ‘we’re spending so much money on travel,'” Megan explained, showing how to add customized metrics that are specific to each client. A profitable client with ten years of runway might swap that widget for gross profit or vendor analysis.

Collaboration happens right on the platform. On any transaction, category, or report, you can leave a question. The client receives a notification and can respond directly from email without logging in. “One of the biggest pain points is transfer of knowledge,” Megan said, “making sure that you have everything that you need from your clients and vice versa.”

Custom reports become interactive stories rather than black-and-white PDFs. The AI generates insights like “You earned 33% more in March compared to the prior month” with drill-down capability to see exactly why. Important insights can be pinned to the executive summary so they’re the first thing clients see.

What this means for your firm

During the Q&A, attendees asked practical questions. One wondered if this integrates with QuickBooks or replaces it entirely. “Digits is a complete ledger system. So it’s a complete replacement,” Megan answered. They can migrate QuickBooks data in about two minutes, but this is a ground-up rebuild, not a bolt-on tool.

Another attendee asked about company scale. The focus is on small and medium-sized businesses, which is the client base most firms serve.

The shift from reviewing everything to reviewing only exceptions makes the close faster and more consistent across your team, less error-prone, and it frees up capacity to serve more clients without hiring proportionally.

“It’s a very exciting time to be an accountant while also a little bit scary,” Megan acknowledged near the session’s end. “I think it’s a time to really lean in and be excited.”

She’s right. The firms embracing AI-native tools now will deliver premium advisory services while their competitors are reconciling bank statements at midnight.

To see these workflows in action, watch the full webinar. Every accountant who signs up gets access to a sandbox demo environment where you can test these workflows with real data. And if you attended live or watch the recording, you can earn CPE credit through the Earmark app. Just search for the course and complete the quiz.

The close is changing. Will you lead that change or follow it?

Not All AI Is Created Equal and Your Next Software Decision Depends on Knowing the Difference

· ·

When Jeff Seibert ran consumer product at Twitter, he asked the finance team for his budget to throw a team event. They said they’d get back to him in 45 days. So he just ran the event without them.

That gap between real-time data and 30-to-90-day delayed financial reports was frustrating, and it eventually led Jeff to build Digits, a new general ledger designed from scratch for the machine learning age. After raising $100 million pre-launch, testing 2,000 monthly closes, and getting 80% of clients closed in under an hour, Digits launched in March 2025. Now, just over a year later, hundreds of accounting firms are onboarding thousands of clients onto the platform.

Jeff launched Twitter’s algorithmic timeline in 2016, and it was one of the first global deployments of machine learning. Now, the AI revolution Jeff helped launch is flooding the accounting profession with claims that are hard to verify. Every accounting software company seems to include AI in its marketing copy, promising everything from “fully automated bookkeeping” to capabilities that don’t add up under scrutiny.

In a recent Earmark webinar, host Blake Oliver and Rob Hamilton, Head of GTM at Digits, pulled back the curtain on how AI in accounting actually works. He was joined by Megan Reid, Product Specialist & Firm Enablement at Digits, who fielded questions throughout the session.

Every AI claim in accounting software isn’t real. But accountants who understand the four core model types (plus one common lie) will make smarter investments, automate the right parts of their workflow, and position their firms for a shift Rob sees coming by the end of 2026.

The AI hype problem (and one question to cut through it)

Before making any technology decision, you need a filter for separating real capabilities from marketing fluff. Rob offered a simple one that cuts through the noise.

He showed screenshots from multiple accounting software companies making bold AI claims. One promised “fully automated bookkeeping.” Another asked, “Do you do AI bookkeeping or do you use a dedicated team of experts?” The positioning has gotten so confusing that firms can’t tell what’s real anymore.

The confusion isn’t new. About five years ago, tech investor Naval Ravikant tweeted, “In most pitch decks, AI stands for Anonymous Indians.” For a long time, that was literally true. Services like Botkeeper rose and fell using offshore labor dressed up as automation. Today, “AI actually means we just bolted on and sent all of your data to ChatGPT,” Rob explained.

Here’s your filter: “AI is the same thing as machine learning,” Rob stated. “If someone is talking to you about AI and they’re not referring to machine learning as the underlying premise, it’s just BS.”

But this filter only works if you understand what machine learning actually is.

Traditional software is straightforward. You write code that tells the computer exactly what to do. It’s tedious to build, but rock solid once it works. Machine learning flips this completely. You feed the system thousands or millions of examples, and the model learns the patterns itself. As Jeff explained in a clip Rob played, “You give the computer the goal state—I want this outcome—and then the computer itself is learning how to do it.”

These models are neural networks. Thousands of hidden layers mimic how neurons connect, based on Google’s 2017 “transformer” research paper (the “T” in GPT). It’s a massive matrix multiplication problem where the system figures out how variables relate to each other.

But machine learning isn’t one thing. Different model types have different strengths and uses in accounting. Understanding these distinctions helps you avoid buying the wrong software and shows you exactly where AI can save time and where vendors are overselling.

The model types that matter (and one that doesn’t)

Rob walked through five categories that get lumped under “AI,” but understanding the differences is what separates informed decisions from expensive mistakes.

Generative models

Large language models (LLMs) are the ones you hear about most, ChatGPT being the prime example. GPT stands for “Generative Pre-Trained Transformer,” and these models generate the most likely continuation of whatever prompt you give them. Rob showed a useful application: turning bullet-point close notes into polished client emails. His advice is to write a “job description” for the AI once. Tell it who it is, give context, specify output format, add examples. Then just paste in different client notes as needed.

But generative models have serious limits. They’re “super eager” and always want to complete prompts, making them prone to hallucinations, or making things up that sound real. They’re bad at math because they generate text rather than calculate numbers. And they’re trained on the internet, not your specific clients. “The ways that it is hallucinating is stuff that maybe even is hard for humans to catch sometimes,” Rob warned. Always review the output.

Agents

These are LLMs with help. You give them a job description, a task, and tools, like computer programs they can use to generate reports, list accounts, or run calculations. The agent makes a plan, uses its tools, checks if the task is done, and loops until complete. Rob showed Digits’ agent answering “I want to hire 20 software engineers next year. Can I afford to?” with a data-backed response.

Guardrails are critical. Microsoft’s early agent “started asking people on dates in the chat,” Rob noted. “You don’t want your accounting agent dispensing dating advice.” Agents work well for updating schedules, running quality checks, and answering analytical questions, but they’re slow and need careful boundaries.

Predictive models

These got Rob visibly excited, and for good reason. These models take an input and predict an output from known options. When the model sees a $5 Starbucks charge, it considers the client’s location, history, and chart of accounts. For a local client, it’s meals and entertainment. For a New York client in California, it’s travel. A $157 Starbucks charge is probably an event, regardless.

What makes predictive models perfect for transaction categorization is they can’t hallucinate; they only choose from existing options. They’re deterministic (same input, same output), include confidence scores, and run fast and cheap once trained.

Digits built a “layer cake” of predictive models:

  1. Client-level (learns each business)
  2. Firm-level (encodes your best practices)
  3. Global (trained on 180 million transactions worth nearly $1 trillion)
  4. An LLM fallback for completely new transactions

The result was over 97% accuracy, compared to standalone LLMs that plateau below 80%, which is about the same as outsourced bookkeepers.

Document extraction models

These combine OCR with layout-aware language models that understand document structure. Previous tools used Amazon’s Mechanical Turk, which relied on humans manually extracting data and took hours. Modern extraction models work in seconds. Digits’ bank reconciliation automatically pulls PDF statements, matches transactions to the exact spot in the PDF, and generates audit reports.

Data analysis

This model is where Rob pulled the rug out. Financial reporting and analysis is “actually just math. It’s not ML.” Computers have done statistical analysis for decades. Could you build an agent to do it? Sure, but it would be slow, expensive, and probably wrong. “If anyone says their AI does reporting and statistical analysis, please ask them what they’re talking about.”

Here’s how the right models map to your month-end close:

  • Book transactions: Predictive models (with LLM fallback)
  • Reconcile statements: Extraction models plus matching algorithms
  • Update schedules: Agents
  • Review and correct: Agents with quality checklists
  • Analyze and report: Statistical analysis plus agents for questions

“Shoehorning an LLM in to solve a problem and just sending a bunch of information is fundamentally incorrect,” Rob emphasized. Each step needs the right model. No single AI approach handles everything.

The 2026 prediction

Understanding model types is just the foundation. The urgency comes from how fast everything is converging.

“Across a large client set in different industry types, it’s highly likely that the month-end close process is looking to be completely automated by the end of 2026,” Rob predicts. Even his “95% automated” hedge probably sounds aggressive. But his logic follows directly from the technology.

If predictive models hit 97% accuracy on transactions, extraction models automate reconciliation in seconds, agents handle schedules and quality control, and statistical analysis covers reporting, then manual work drops to a fraction. Rob’s goal is to see accountants doing “1/20th of the work you’re doing today.”

He acknowledged limits. Construction firms with complex job costing might not hit that threshold. But for firms serving professional services, cash-basis businesses, and straightforward accrual clients, the automation curve is steep.

An AI-native firm will focus on value instead of tedium. Deeper industry expertise. Stronger client relationships. Higher margins. You’re not reviewing every transaction, you’re supervising the system and handling exceptions. Those hours saved on reconciliation become advisory time clients actually value.

But this is also a competitive necessity. “AI won’t replace you. Someone who’s good at using AI is going to,” Rob said, quoting a common warning. And he was direct about the stakes. Firms that don’t adapt will face “a cascading effect on business models” as early adopters pull ahead.

For overwhelmed firms, which Rob acknowledged includes most firms, he offered practical starting points:

  • Map your processes first. If you use workflow tools like Karbon or Keeper, you’ve probably documented your steps. If not, start there. You can’t identify where AI fits until you know what you’re actually doing.
  • Start small and low-stakes. Don’t tackle your biggest challenge first. Try drafting emails, testing categorization, or visualizing data. Build your intuition gradually.
  • Get hands-on with new tools. Rob mentioned being impressed by Claude Opus, which could build HTML dashboards from his data (something he couldn’t do as a non-engineer). The specific tool doesn’t matter; hands-on experience builds judgment.
  • Know your business before choosing where to start. As Rob put it, “You need to know the details of your business to know where you can start and where the right places to poke and prod are.”

The “wait and see” window is closing. Firms that develop AI literacy now by asking questions about models, data handling, and use cases, will be ready for the rest of 2026 and beyond.

Your next move: Better questions, smaller steps, faster action

Let’s turn Rob and Megan’s insights into actionable takeaways:

  • Not all AI is equal. Four real model types plus one fake (data analysis) get lumped together. When vendors pitch “AI-powered reporting,” you now know to dig deeper.
  • Each close step needs a different model. Predictive for transactions. Extraction for reconciliation. Agents for schedules. Statistics for reporting. Anyone claiming one solution does everything deserves scrutiny.
  • Predictive models beat LLMs for categorization. Layered architectures that learn your clients and firm patterns dramatically outperform chatbots. Bigger isn’t always better.
  • Ask vendors the hard questions. What model type? Where does data go? Are you training your own models or sending financial data to third parties? This is due diligence.
  • The tipping point is closer than you think. Whether Rob’s 2026 prediction proves exactly right or directionally right, the trajectory is clear. Understanding these distinctions now positions you to take action.

For those interested in going deeper, Rob mentioned resources like the AI Native Accounting Foundation and the AI-Native Accounting podcast hosted by Kacee Johnson, where industry leaders discuss the latest developments.

The accounting profession is at a real inflection point. Smart firm leaders will develop the literacy to ask smart questions, experiment in the right places, and redirect time from tedium to advisory work clients value.

Rob noted this might be one of the few professions with such clear AI use cases, putting accountants at the forefront of innovation. That’s an opportunity to shape how technology serves the profession, not the other way around.

Watch the on-demand webinar for complete details, including live demonstrations, security architecture specifics, and audience Q&A covering nonprofits, inventory clients, and platform migrations. The future of accounting is being written now. Make sure you’re part of the conversation.

Four Years of Cleanup in Four Hours: Inside the AI Ledger That Learns From Your Work

· ·

What if you could stop programming bank rules forever? No more tweaking text strings, adding exceptions, or debugging why “COSTCO WHSE #1234” won’t match your Costco rule. During a recent Earmark Expo webinar, accounting software company Digits demonstrated exactly how that future works, and they achieve 96% automatic categorization accuracy without a single bank rule.

Host David Leary has been watching Digits since before ChatGPT existed. “I remember seeing a pitch deck about Digits, and it was being emailed around on backchannels in the accounting industry,” he recalled. “This pitch deck was super ambitious. At the time, the back channel hallway talk was like, ‘Great, here comes another bank feeds accounting app.’”

Now, years later, that ambitious vision is reality. Rob Hamilton from Digits’ partnerships team showed David and his co-host Blake Oliver what the company calls the world’s first “agentic general ledger,” software built from the ground up with machine learning at its core.

Why Digits Took Six Years to Build

Before diving into the technology, Rob shared the origin story. Digits founder Jeff Seibert sold his previous companies to Box and Twitter. At both companies, he noticed a stark contrast: product and engineering teams had real-time dashboards showing exactly who was on their website and what buttons they clicked. But when he wanted to check if he had a budget for a team event, finance told him to wait 45 days for the books to close.

“As a founder of a company, you’re like, ‘This is crazy. I’m just going to do this event without your approval,’” Rob explained. When Jeff left Twitter, he wanted to use machine learning for good, and accounting emerged as the perfect candidate.

The result took six years to build. “Turns out that it takes a while to build a general ledger from the ground up in the machine learning era,” Rob admitted. But that ground-up approach makes all the difference.

The Three-Layer Intelligence That Replaces Bank Rules

Traditional accounting software makes you act like a programmer. You write rules, define patterns, and hope the software follows instructions. Anyone who’s debugged bank rules knows the frustration.

Digits flips this completely. Instead of you teaching the software through rules, the system learns from your work at three levels.

First, it learns from each specific company. When you connect QuickBooks to Digits, it imports your historical data and trains on how you categorize that company’s transactions. “We actually train on a company level,” Rob explained. “When transactions start coming in, it actually leverages the work you’ve already done within that individual company.”

Second, it learns from your entire firm. When a new vendor appears—say, a coffee shop that just opened—Digits checks if any other client in your firm has seen that vendor. Your work for one client helps all your clients.

Third, it taps into global intelligence. For truly novel transactions, Digits uses its global model trained on every transaction the platform has ever processed.

The payoff is significant. “For September, we’re at a 96% rate of transactions getting booked into Digits that then were subsequently not touched by a human afterwards,” Rob revealed. That’s not just categorized; that’s categorized correctly enough that accountants didn’t change them.

“You’re not editing any rules,” Rob points out, contrasting Digits with traditional systems. “You don’t have to add an extra appendage to pull out the specific Costco transaction. We learn from your behaviors directly inside the product.”

How the System Handles the Other 4%

No AI system is perfect. What matters is how it handles uncertainty. When Digits encounters a transaction it’s unsure about, it doesn’t guess silently. It flags the uncertainty and shows its reasoning.

During the demo, Rob showed a US Patent and Trademark Office transaction where Digits displayed, “I have this as taxes, but I actually think it could be legal.” The system even suggested adding “intangibles” as a new account category for companies still building their chart of accounts.

The learning happens instantly. “We’ve built our architecture to be uniquely quick in the training,” Rob emphasized. “The second we see a similar transaction, it’ll effectively be perfect based on your prior action.”

Quality control is proactive rather than reactive. Each month, Digits flags all new vendors so accountants can verify they’re categorized correctly. It also highlights vendors booked to multiple categories, like Apple transactions split between fixed assets and software subscriptions.

When accountants don’t know what something is, they can ask clients directly within the platform. Questions attach to specific transactions, clients get email notifications, and responses flow back to the same transaction. The AI suggests categorization based on the client’s answer, though accountants confirm before applying.

David appreciated the unified workflow. “Now I don’t have to have five browser tabs open where one browser tab is the report, the transaction is in a new browser tab, and I make the edit and refresh the report in the other browser tab.”

Reconciliation in Minutes, Not Hours

Bank reconciliation should be simple, but when something doesn’t match, like $15 missing from Stripe, the detective work begins. Digits transforms this process entirely.

Statements enter the system three ways. Banks like Wells Fargo send them automatically via API. For others, accountants drag and drop PDFs directly onto the platform. Every Digits account also gets a single email address that accepts any document type, including statements, bills, or receipts, The AI routes them appropriately.

“So one email for all the transactions in a client’s company file,” David noted. “You don’t have special HR email and AP email where you send it to the wrong box and it creates a mess.”

The reconciliation interface shows the bank statement PDF alongside ledger transactions. As you hover over transactions, green boxes highlight the matching line on the statement. David’s reaction captured what every accountant will recognize: “I used to do this with a highlighter and my fingers. I had to find it on both.”

But the real magic is proactive problem detection. Digits identifies specific issues and offers one-click fixes for things like:

  • Uncleared transactions that should move to next month
  • Statement items missing from the ledger
  • Date discrepancies between records and statements

Each issue comes with a resolution button. The system does the detective work; accountants just confirm the fix.

“We had an accountant come in the other day. He was like, ‘I did four years of cleanup in four hours’ because he just linked the bank accounts, dragged all the statements in, and the AI did everything,” Rob says.

Beyond Bookkeeping: Reports Clients Actually Read

With traditional financial reports, only 15% of business owners even open those black-and-white PDF attachments. Digits studied this and found that when firms use visual reporting tools, over 70% of clients actually open and interact with the financials.

The reporting system works like “Google Docs for your finances,” as Rob described it. Accountants can add commentary directly on line items, tag clients with questions, and create visual dashboards that tell the story of the business.

The platform includes built-in bill pay ($0.50 for ACH, $2 for checks) and invoicing. The system automatically recognizes and routes dragged-in documents. Bills queue for payment, receipts match to transactions, and statements trigger reconciliation.

Behind the scenes, AI agents continuously research every vendor, building what Rob called “a dossier” with logos, phone numbers, and descriptions. “This is what your team does when they don’t know what a transaction is. They Google it and find the information.”

What This Means for Your Practice

The shift from rule-based to AI-native software fundamentally changes the accountant’s daily work. Instead of programming rules, you review AI suggestions. Instead of hunting for reconciliation errors, you confirm one-click fixes. Instead of sending reports that get ignored, you create interactive dashboards that clients actually use.

The compound effect is striking. Every correction teaches the system, improving accuracy for that client, your entire firm, and eventually all Digits users. Time savings stack up, allowing firms to shift toward advisory work.

Digits offers a partner program with volume discounts. The standard price is $100 per month per client for full features, with special pricing for tax write-up work. Accounting firms get their own firm account free when joining the partner program.

Rob emphasized that construction and other complex industries might see slightly lower accuracy rates than the 96% average, but the system continuously learns and improves. Features like sales tax support and project tracking are coming soon, while departments and locations tracking are already available.

For firms evaluating new software, the question has shifted from “What rules do I need to create?” to “How well does this system learn?” The four-years-in-four-hours cleanup example shows what’s possible when AI handles the tedious work.

Watch the complete Earmark Expo webinar to see the full demonstration, including reconciliation workflows, client communication tools, and the visual reporting system that gets clients actually engaging with their financials. Whether you’re ready to switch or just want to understand where accounting technology is heading, this demo shows what accounting looks like when bank rules become obsolete.