Write-Up Work in 2026: Why Most Tax Preparers Still Use Excel (And What's Changing)

Tax season hasn't changed much in 20 years. Not the core of it, anyway.

Clients still show up in January and February with a pile of bank statements, some missing, some for accounts you didn't know they had. You still open Excel, create columns for date, description, amount, and category. You still scroll through hundreds of transactions, one at a time, deciding where each one goes on Schedule C or whatever form you're preparing.

And you still do this for every single client, every single year, knowing that half your time is spent on transactions that are obvious (yes, the Verizon bill is a utility) and the other half is spent on transactions that require actual thought (is this Home Depot charge supplies, equipment, or a personal project?).

This is write-up work. It's the foundation of every tax engagement for clients who don't use a bookkeeper, which in practice means most sole proprietors, many partnerships, and plenty of small S-corps. And it is, by a wide margin, the most time-consuming part of preparing a small business return.

Why Excel Has Been the Default

Excel works. That's the simplest explanation.

A preparer opens a CSV export from the client's bank, adds a category column, and starts typing. No software to learn. No subscription to pay. No integration to configure. The barrier to entry is zero, and the workflow is familiar to anyone who has used a computer.

For a preparer with 15-20 write-up clients, Excel is manageable. It takes 2-3 hours per client, which adds up to maybe 40-50 hours of write-up time during tax season. That's a lot, but it's predictable, and the preparer maintains complete control over every categorization decision.

Excel also has no opinions. It doesn't try to auto-categorize anything, which means it never gets anything wrong. The flip side is that it never gets anything right either. Every transaction requires a manual decision, even the ones that are identical to transactions you categorized last year for the same client.

Where Excel Breaks Down

The problems start when volume increases. A preparer who grows from 20 write-up clients to 50 or 100 hits a wall. At 50 clients with an average of 800 transactions each, you're looking at 40,000 categorization decisions. Even at 10 seconds per transaction (which is fast), that's 111 hours of pure categorization, not counting the time to request statements, chase missing months, reconcile, generate the P&L, and enter the numbers into tax software.

Excel also doesn't learn. If you categorize "SYSCO FOODS" as Cost of Goods Sold for a restaurant client in January, and the same vendor appears 47 more times on the same statement, you're making the same decision 47 times. If you process three restaurant clients, you're making it 141 times. Excel doesn't remember that you already decided SYSCO is COGS.

The other weakness is that Excel categories are whatever you type. There's no structure forcing you to use actual tax return line items. One preparer might type "Meals," another might type "Meals & Entertainment," and a third might type "M&E 50%." They might all mean Schedule C Line 24b, but when it's time to enter the numbers into tax software, you still have to translate your freeform categories into the specific lines on the return.

What's Actually Changed in 2026

Two things have changed that make this the right time to rethink write-up workflows.

The first is AI. Large language models can now read a transaction description like "AMZN MKTP US2K7X9" and understand that it's an Amazon Marketplace purchase. They can look at "SQ MARIO'S PIZZA" and know it's a restaurant (Square terminal, meals category). They can read "GUSTO 032026 PAYROLL" and flag it as a payroll processor withdrawal that likely needs to be split across wages, payroll taxes, and fees.

This isn't hypothetical. The technology works today. The question is how you use it. Running every transaction through an AI model one at a time is slow and expensive. The better approach is rules-based matching first (use a database of known vendors and their categories), then batch the remaining unknowns through AI for suggestions, then flag anything the system isn't confident about for human review. That way, you get speed on the easy ones and intelligence on the hard ones, without giving up control.

The second change is the OBBBA (One Big Beautiful Bill Act), signed July 4, 2025. The new law made several changes that affect how preparers approach categorization:

100% bonus depreciation is permanently restored for qualifying property acquired after January 19, 2025. This means flagging large equipment purchases is more important than ever, because the full cost can be deducted in year one. Getting these transactions categorized correctly (on Form 4562 via Line 13, not buried in Supplies on Line 22) matters for audit defense even if the tax result is the same.

Section 179 limits jumped to $2.5 million (inflation-adjusted to $2.56 million for 2026) with a phase-out threshold starting at $4.09 million. Small businesses can now expense virtually any equipment purchase immediately.

Employer-provided meals for the convenience of the employer are no longer deductible starting in 2026. Through 2025, these were 50% deductible. Now they're 0%. This affects how you categorize on-site cafeteria expenses, break room meals provided during overtime, and similar employer-provided food. Business meals with clients remain 50% deductible.

What a Modern Write-Up Workflow Looks Like

The ideal write-up workflow in 2026 keeps the preparer in control while eliminating the repetitive work.

Step one: the client uploads their bank statement CSV (or the preparer downloads it from the bank). No manual data entry, no retyping transactions from paper statements.

Step two: the system matches transactions against a vendor rules database. If "VERIZON WIRELESS" has been categorized as Utilities on the last 200 times it appeared across all your clients, it gets auto-categorized as Utilities. If "STATE FARM" was categorized as Insurance for this specific client last year, it gets the same treatment. These rule-based matches are instant, free, and as accurate as the rules you've trained.

Step three: transactions that don't match any rule go to AI for a suggested category. The AI looks at the vendor name, the amount, the client's industry, and the transaction pattern to make its best guess. It provides a confidence score and reasoning: "Categorized as Line 24b: Meals (50% deductible) because vendor 'PANERA BREAD' is a restaurant chain."

Step four: anything the system isn't confident about gets flagged for the preparer's review. Amazon purchases, Venmo payments, large one-time charges, anything that could go multiple ways. The preparer makes the final call, and that decision feeds back into the rules database so the system learns.

Step five: the output is a P&L mapped directly to the tax form's line items. Not generic accounting categories that need translation, but the actual lines on Schedule C, 1065, 1120S, or 1120. You can enter the numbers into your tax software without an intermediate step.

The result: what used to take 2-3 hours takes 15-20 minutes. The preparer's expertise goes where it matters (the ambiguous transactions, the judgment calls, the client conversations) instead of being spent on the 80% of transactions that are obvious.

What This Means for Your Practice

Write-up work isn't going away. As long as small business owners don't use bookkeepers (and most don't), someone needs to turn bank statements into P&Ls. The preparers who figure out how to do this efficiently will handle more clients, deliver faster turnaround, and spend their expertise where it actually affects the tax return.

The preparers who stick with pure Excel will keep doing fine with a small client base. But if you're looking to grow, or if you're already drowning in write-up volume during January through April, the math is pretty simple. At 2 hours per client, 50 clients is 100 hours of write-up work. At 20 minutes per client, the same 50 clients take about 17 hours. That's 83 hours back, and every one of those hours is during your busiest season.

The shift isn't about replacing preparers. It's about moving expertise to the decisions that actually require it — the judgment calls, the client conversations, the OBBBA-driven category changes — instead of spending it on the 80% of transactions that were never ambiguous in the first place.