Last spring, a homeowner in Fort Bend County, Texas, received a property tax assessment valuing her 2,400-square-foot ranch at $412,000, roughly $47,000 above what three recent comparable sales within half a mile supported. She did not protest the assessment, and neither did 71 percent of her neighbors. Across Harris County alone, 386,871 homeowners who could have protested chose not to, leaving a collective $248 million in potential savings unclaimed.
She is not unusual. She is the median American homeowner.
A $339.8 Billion System Nobody Audits
Americans paid $339.8 billion in residential property taxes last year. I have spent enough years watching construction budgets bleed from unchecked assumptions to recognize the same pattern in a different ledger: a system so large that its inefficiencies become invisible, accepted as fixed costs rather than addressable errors. Property tax assessment is that system, and the inefficiency is staggering in both scale and predictability.
According to Realtor.com's 2025 analysis, 40.5 percent of U.S. properties may be overassessed relative to their actual market value, while data from the National Taxpayers Union and Bankrate shows that only 3 to 5 percent of homeowners file a formal protest, and of those who do, between 30 and 50 percent win a reduction.
Run those numbers against each other slowly and a gap emerges that should alarm anyone who has ever scrutinized a tax bill, a construction change order, or any other document where someone else decided what you owe and counted on you not checking. If 40.5 percent of homes are overassessed and only 4 percent of homeowners appeal, approximately 36 percent of all residential property owners are overpaying without contest. At $539 median annual savings per successful appeal, applied across roughly 90 million owner-occupied homes in the United States, the aggregate unclaimed savings exceed $17 billion annually. Every year. That is a structural transfer of wealth from inattentive homeowners to county tax rolls, and it has been happening for as long as mass appraisal systems have existed, quietly and without protest, because the people losing money do not realize they are losing it.
Why Nobody Protests (and Who Profits from the Silence)
An Ownwell survey of 2,500 homeowners in 2025 found that 53 percent were unaware they had the right to appeal their property tax assessment at all, not because they chose to accept the number but because they did not know the option existed. Another 74 percent reported anxiety about rising property taxes but had taken no action, which is the fiscal equivalent of watching water pool on your foundation slab and deciding to worry about it next season instead of walking outside with a level and a shovel.
Filing a protest is free. In most jurisdictions, the process involves submitting a one-page form to your county appraisal district with comparable sales data supporting a lower valuation, no attorney required, no filing fee, and no risk of your assessment increasing as retaliation, a common misconception that keeps millions of homeowners compliant. You fill out the form, you show up at a hearing or submit evidence online, and a review board decides whether your neighbor's house that sold for $380,000 is a better comparable than the one the assessor picked three miles away in a different school district.
It is free, and almost nobody does it.
Enter the Algorithms
Ownwell, a Texas-based startup that raised $50 million in Series B funding in February 2026 and has processed over one million appeals to date, represents the most visible entrant in a growing category of AI-powered property tax appeal platforms. Their system ingests millions of local property records, identifies overassessed homes, generates appeal packets with comparable sales analysis, and in many jurisdictions handles the entire protest process without the homeowner attending a hearing or filling out a single form.
Their numbers are worth examining carefully, not because I expect them to be fabricated but because the difference between a headline success rate and a typical homeowner's experience is where the useful information lives. Across their primary service areas in Texas, Illinois, Florida, Georgia, California, Washington, and New York, Ownwell reports an 86 percent success rate on appeals, with average annual savings between $774 and $1,100 per customer. In Fort Bend County, where 29 percent of homeowners protested, 86.5 percent of protests succeeded, producing $194 million in collective reductions. Travis County hit 88.4 percent success on a 41.5 percent protest rate, the highest participation in the state, suggesting that communities where protest culture is normalized see compounding benefits as neighbors share what worked and new filers enter the process with better-prepared evidence packets.
Ownwell charges nothing upfront, collecting instead a fee of 25 percent of first-year savings, which means a homeowner who saves $800 pays $200 and keeps $600 that would have otherwise gone to the county. In year two and beyond, the full reduction carries forward at no additional cost unless the homeowner uses the service again for a new protest cycle, which means the real value proposition is not the first-year savings, discounted by 25 percent, but the cumulative reduction compounding silently through every subsequent tax year.
In Texas alone, Ownwell managed appeals for more than 200,000 properties where owners had never previously protested, a number large enough to shift county revenue projections and force assessor offices to reconsider whether their valuation models can withstand organized, data-driven challenge at that scale. Two hundred thousand homeowners who had been silently overpaying, year after year, until an algorithm noticed what they had not. Scale matters here.
The Equity Problem Nobody Wants to Quantify
If overassessment affected all homeowners equally, this would be a simple efficiency story. It does not.
Research from the University of Chicago Harris School of Public Policy found that 97.7 percent of U.S. counties have regressive property tax assessments, meaning cheaper homes are systematically assessed at higher effective rates than expensive ones. Their average 10/90 ratio, which compares the assessment rate on the cheapest 10 percent of homes to the most expensive 10 percent, is 133 percent. Nearly half of all counties have ratios above 150 percent, meaning the cheapest homes in those jurisdictions are assessed at one and a half times the rate of the most expensive ones.
A Federal Reserve Bank of Philadelphia working paper put a finer point on it: owners of inexpensive houses pay approximately 50 percent higher effective property tax rates than owners of expensive houses. Sixty percent of that regressivity comes from flawed valuation methods that ignore variation between houses and neighborhoods, and the remaining forty percent comes from infrequent reappraisal, a compounding problem in markets where home values have shifted dramatically since the last assessment cycle. A simple improvement to the valuation methodology, the Fed researchers calculated, could increase poor homeowners' net worth by more than 10 percent, a figure that reframes property tax reform from a budgetary nuisance into one of the most accessible wealth-building interventions available to low-income homeowners in the country.
Layer the protest data on top of the regressivity findings and the picture shifts from troubling to genuinely difficult to defend as anything other than a systemic equity failure. Rice University's Kinder Institute analyzed Harris County property tax protest patterns and found that predominantly Black neighborhoods were 40 percent less likely to file a protest than white neighborhoods, and Latino neighborhoods were 33 percent less likely. When residents of those neighborhoods did protest, their appeals were 36 percent more likely to result in no change to the assessment, meaning the system penalizes them twice: once for the inflated valuation and again for the lower probability of successful appeal when they do challenge it.
So the system overassesses cheap homes at higher rates, the communities most affected by that overassessment are least likely to protest, and when they do protest they are least likely to win. AI appeal platforms could theoretically democratize access to a process that currently rewards the informed and the persistent, but whether they actually reach the homeowners who need them most is a question that none of the companies in this space have answered with data. Not one.
The Counterargument That the Platforms Would Rather You Not Hear
Ownwell and its competitors charge 25 percent of savings for automating a process that is free to do yourself. That fee structure profits from information asymmetry. Pure and simple. Specifically, it monetizes the gap between homeowners who know they can protest for free and homeowners who do not. Rice University's Harris County data makes the counterargument sharply: DIY homeowners who protested without hired help achieved a median reduction of $7,889, compared to $5,125 for homeowners who used a paid agent. Homeowners who did the work themselves, pulling comps, printing forms, showing up to the hearing, saved 54 percent more than those who outsourced the process to a company that promised to handle everything.
There is also the matter of a Lanham Act lawsuit filed by Texas Tax Protest against Ownwell, raising questions about the accuracy of the company's marketing claims regarding success rates and savings figures. When a competitor sues you over your numbers, it does not prove the numbers are wrong, but it does suggest that someone with financial incentive to check found enough ambiguity to file a federal complaint.
I have watched this dynamic play out in construction technology for twenty years, always following the same script, and what follows is so predictable it barely qualifies as analysis. A company identifies a real inefficiency, builds a tool that genuinely helps, and then monetizes the gap between what people know and what they could learn for free. Ownwell's 86 percent success rate is real. Meaningful. But 86 percent of what, exactly? If the platform cherry-picks properties with the clearest overassessment signals, declining to file appeals where success is uncertain, the headline number overstates the typical homeowner's expected outcome. Ownwell has not published case-level data that would let us evaluate selection effects.
Until they do, treat the 86 percent figure as a marketing claim with strong supporting evidence, an asterisk the size of a property line dispute, and the understanding that your experience may not match the headline.
What You Should Actually Do
If you own a home and have not reviewed your property tax assessment in the last two years, stop reading this and go pull your county's assessment notice. Compare the assessed value to three to five recent comparable sales within half a mile of your property, matched for square footage, lot size, year built, and condition. Zillow, Redfin, and your county assessor's website all provide this data for free, and any one of them will give you a clearer picture than whatever automated valuation model the assessor's office ran two years ago when they last recalculated your number. If your assessed value exceeds the median of your comparables by more than 5 percent, file. You have a case.
Filing costs nothing, and the worst outcome is that the board agrees with the current assessment and nothing changes. Your assessment cannot increase as a result of your protest in the vast majority of jurisdictions, and the process takes between two and six hours of total effort depending on whether your county offers an online submission or requires an in-person hearing.
If you do not want to do the research yourself, Ownwell and similar platforms like Texas Tax Protest, Kroll, and various local firms will handle it for a contingency fee. At 25 percent of savings, the math works if your expected reduction exceeds roughly $400, because you will net $300 or more after fees, and the time investment drops to approximately fifteen minutes of signing authorization forms.
| Approach | Cost | Time Investment | Median Savings (Harris County) | Net After Fees |
|---|---|---|---|---|
| DIY protest | $0 | 2-6 hours | $7,889 | $7,889 |
| AI platform (Ownwell) | 25% of savings | 15 minutes | $5,125* | $3,844 |
| Traditional tax agent | 30-40% of savings | 30 minutes | $5,125* | $3,075-$3,588 |
| No action | $0 | 0 | $0 | -$539/yr (median overpayment) |
*Paid-agent median from Rice University Kinder Institute Harris County analysis. Savings vary by jurisdiction and property characteristics.
Where This Is Headed
Ownwell's partnership with Realtor.com, which integrated appeal tools directly into the My Home dashboard, signals where the market is moving: property tax protest as a default feature of homeownership platforms rather than a specialized service you have to seek out. When the appeal button sits next to your Zestimate and your mortgage balance, the 53 percent of homeowners who did not know they could protest will find out by accident. That is probably the most effective distribution strategy anyone has devised for this category, and it bypasses the equity problem entirely by meeting homeowners where they already are rather than requiring them to search for a solution they do not know exists.
Assessor offices are not standing still on the other side of this equation, either. Several counties have started deploying their own AI tools to improve assessment accuracy on the front end, which would reduce the number of overassessments before they reach homeowners, shrinking the appeal market before commercial platforms can fully entrench themselves. If the assessment improves, the appeal opportunity shrinks. Platforms lose their market. Obvious. Whether county governments will invest in assessment technology at the pace required to close the gap before commercial platforms entrench themselves is an open question. Based on twenty years of watching municipal technology procurement, I would not bet on speed. Never do.
Limitations
This analysis relies on Ownwell's self-reported success rates and savings figures, which have not been independently audited and are the subject of active litigation. Realtor.com's 40.5 percent overassessment estimate uses a proprietary methodology that has not been peer-reviewed. Rice University's Harris County protest data covers one county in one state with an unusually active protest culture, and the DIY-vs.-agent savings gap may not replicate in jurisdictions with different hearing procedures. We could not verify whether Ownwell's 86 percent success rate reflects all filed appeals or only appeals that Ownwell chose to pursue after screening, a distinction that materially affects interpretation. Federal Reserve Bank of Philadelphia's regressivity findings used national data that masks significant variation between states with different assessment and equalization practices.