Last spring, the Peoria County assessor decided my little machine shop was suddenly worth 40% more than it had been in October. No explanation. No fanfare. Just a number in an envelope that made me choke on my coffee.

I hired Dave Kowalski to fight it. Dave’s sharp. He knows the appeals process better than I know my own toolbox. But when he showed up to review my file, he looked like he’d lost a bar fight with a filing cabinet. Six hours at my desk doing the same thing over and over: squint at a scanned PDF tax bill, type a number into Excel, squint again, delete it, retype it. By hour three, I wanted to pour him a whiskey. By hour four, I realized this guy was spending maybe a third of his life as a very expensive copy-paste machine.

That stuck with me. I spent the next few weeks calling every property tax consultant I know in central Illinois to see if they were living like this too.

They were.

Tom Hess, who runs Hess Property Advisors down in Morton, told me he spent his last three Sundays manually typing rent rolls into Excel because his intern quit. Angela Voss at Voss & Associates in Bloomington said her firm burns through three paralegals a year, mostly because the job is 80% data entry and 20% actual thinking. And Dave—well, Dave was still using reading glasses and a highlighter.

So I demoed six of the most talked-about platforms. Most were disappointing. One startup, which I’ll keep nameless because their lawyers probably dress better than I do, charged me $200 just to see a dashboard that looked like a high school coding project. The login worked great. The AI couldn’t read a Tazewell County tax bill without inventing a PIN number.

But a few were worth the trouble. Here’s the honest breakdown, with real prices and actual opinions.

Levanta: Enterprise Only, Enterprise Price

Levanta is the only tool here built specifically for tax appeal workflows from the ground up. It pulls property data straight from county assessor feeds, runs valuation models automatically, flags inequities, and produces complete appeal packages.

Marcus Chen, who runs Equity Tax Group up in Chicago, swears by it now. He told me it cut his prep time by about two-thirds once it was running. But—and this is a big but—he admitted the first three months were a nightmare. Levanta needs clean, structured data going in. Feed it garbage, get fast garbage.

The pricing is enterprise-only: $20,000 to $100,000-plus per year depending on portfolio size and API volume. There’s no credit card signup. Implementation takes months.

For a solo practitioner like Dave, or even Angela’s ten-person shop in Bloomington? Absurd. Dave handles maybe forty appeals a year. Marcus handles four thousand. The math only works at scale.

Docsumo: It Reads the Bad PDFs So You Don’t Have To

Docsumo uses OCR and machine learning to pull numbers out of scanned tax bills, rent rolls, and financial statements. It turns blurry PDFs into clean spreadsheet rows.

Dave pays $300 a month for the Starter plan, which covers 1,000 pages. Growth is $600 for 3,000 pages.

Here’s the math that matters: before Docsumo, Dave paid an intern $18 an hour to type numbers. The kid worked six hours a week. That’s $432 a month, plus payroll taxes, plus the time Dave spent checking the work because teenagers don’t know a PIN from a parcel number. Docsumo doesn’t call in sick, doesn’t need workers comp, and after Dave trained it on his specific document formats, it hits about 95% accuracy on complex tables.

The catch? You can’t upload a PDF day one and expect magic. Dave said the setup was tedious. “It’s not magic. You have to show it maybe twenty documents before it stops trying to read the watermark as a dollar amount.” And if you’re processing massive volume, that bill scales linearly.

But compared to watching another human type numbers until their eyes bleed? For any consultant drowning in paper, Docsumo is a no-brainer.

ChatGPT: Cheap Help That Hallucinates

ChatGPT Team costs $25 per user per month billed annually, or $30 month-to-month. Enterprise is custom.

For that money, you get a drafting assistant that actually understands tax appeal language. Dave uses it to generate first drafts of jurisdiction-specific letters. He’ll upload a spreadsheet of comps, ask it to find outliers, and have it rewrite client emails so they don’t sound like a lawyer having a stroke.

Angela’s team dumps huge case documents into it and asks it to find weak spots in their argument. It’ll build pivot tables from a CSV. It’ll summarize a 50-page lease in about thirty seconds.

But—and this is a big, flashing, neon but—it hallucinates. It will absolutely invent case law. Tom told me it once cited a “landmark” Illinois circuit court decision that turned out to be completely fictional. “Sounded real authoritative, though,” he said. You have to verify every citation.

And if you’re not on Enterprise, scrub your client data before uploading anything. Don’t feed it Social Security numbers or parcel records unless you want to explain a data breach to the Illinois Department of Financial and Professional Regulation.

Thomson Reuters built CoCounsel on top of Westlaw, which means it checks statutes and pulls real case law instead of making it up.

It’s $99 a month for an individual plan. The good stuff usually gets bundled into Westlaw subscriptions that run $500-plus a month.

Tom uses it for complex litigation and multi-jurisdictional research. He said it summarized a 100-page contract in the time it took him to walk to his kitchen and back. Another consultant I know said it cut her research time in half on a Cook County commercial dispute.

But Tom also told me the interface felt like “teaching my mom to use Photoshop.” Steep learning curve. Lots of features you’ll never touch.

For Dave, who mostly needs to look up Illinois assessment ratios once a quarter? CoCounsel is a sledgehammer on a thumbtack. For Angela, who has three active lawsuits and properties in six states? It’s worth every penny of that Westlaw bundle.

Copilot: Already in Your Office, Already Boring

Microsoft Copilot shoves AI into Excel, Outlook, and Word for $30 per user per month on top of your existing Microsoft 365 license. Since you’re already paying $36 to $57 a month for that, your real total is $66 to $87 per user.

Angela’s firm pays for it because her team lives in Excel building DCF models. Copilot writes complex formulas, cleans dirty data, formats reports, and drafts client emails. Nothing leaks outside Microsoft’s servers. That matters if you’re paranoid about client data—and you should be.

The downside? It’s generic. It won’t pull comps from the county assessor. It chokes on huge datasets; Angela said a 100,000-row spreadsheet made it gasp like a fish on a dock. And that total cost adds up fast. You’re not just buying Copilot; you’re buying the whole Microsoft ecosystem to prop it up.

Dave tried it for a month and canceled. “It writes great emails,” he said. “But I don’t need great emails. I need someone to read my PDFs.”

HouseCanary: Residential Gold, Commercial Garbage

HouseCanary serves up AI-driven valuations, market forecasts, and comparable sales data. Portal access starts around $500 a month. API access runs $5,000 to $10,000 a month depending on call volume.

Angela uses it for her residential portfolio. The comps data is solid, and the API auto-updates valuations across hundreds of properties without her team clicking through county websites.

But for commercial work? The data gets thin. Dave tried it on a warehouse appeal in Pekin and said the “comparables” were a convenience store and a church. Like bringing a butter knife to a steak dinner.

Plus, API integration requires a developer. If your IT department is a nephew who “knows computers,” you’re stuck with the portal. For a small shop without a tech person, that’s a problem.

What You Should Actually Buy

If you’re Dave—solo shop, mixed commercial and residential, forty appeals a year—you don’t need Levanta’s enterprise price tag or HouseCanary’s residential bias. You need Docsumo ($300) to kill the data entry, ChatGPT Team ($25) to draft letters, and a working knowledge of Illinois statutes. Total: $325 a month. That’s less than the intern he fired, and it doesn’t request time off to study for the LSAT.

If you’re Angela—mid-size firm, multi-state, heavy residential—you’re looking at Docsumo Growth ($600), ChatGPT Enterprise for privacy, CoCounsel through Westlaw ($500-plus), HouseCanary ($500), and maybe Copilot ($30 each for five users). She’s in for about $3,000 to $5,000 monthly. Still cheaper than two junior analysts and a paralegal, which is what she’d need otherwise.

If you’re Marcus in Chicago with thousands of appeals? Levanta finally makes sense. At $50,000-plus a year, it’s a monster investment. But at that volume, automation pays for itself in the first quarter.

The Bottom Line

For most consultants reading this—guys running small shops in Peoria or Springfield or Decatur—start small. Get Docsumo. Add ChatGPT Team. Prove the ROI for three months. Then worry about the fancy stuff.

The goal isn’t to impress people at industry conferences with your tech stack. The goal is to finish your appeals by five o’clock, have dinner with your family, and not need reading glasses at age forty because you spent another night squinting at a scanned PDF from Macoupin County.

Dave’s been home by six every night since February. He says his blood pressure’s down. And my machine shop’s assessment got reduced, by the way.

Sometimes the old ways work fine. Sometimes you just need a robot to read the paperwork first.

Meta description: A property tax consultant tests six AI tools for tax appeal work—Docsumo, ChatGPT, CoCounsel, Copilot, HouseCanary, and Levanta—with real prices and honest opinions on what’s worth buying for small shops.

Related reading:

Top AI Tools for Real Estate And Property

Here are the tools we recommend for real estate & property:

Disclosure: Some links in this article are affiliate links. We may earn a commission at no extra cost to you.

About the author: Kai is an AI tools analyst who tests software so you don’t have to. He runs a small business in Peoria, IL and writes about the tools that actually save time and money.