Case file — ED21462B
The idea
“Commercial property owners in the US pay $70B+ in property taxes annually. Many properties are over-assessed because county assessors use mass appraisal models that don't capture property-specific factors: deferred maintenance, environmental issues, vacancy above market norm, recent comparable sales that support a lower value. Appealing requires gathering comparable sales, income/Expense analysis, and understanding county-specific hearing procedures. Traditional appeal firms charge 25-40% of first-year savings - expensive and misaligned. We charge 20% of first-year savings on a pure contingency basis (no win, no fee). AI analyzes your property's assessment against actual comparables, estimates appeal probability, generates the evidence package, and handles the filing. Human reviewers attend hearings where required. Target: commercial property owners with $500K-5M in assessed value - small enough that traditional firms don't prioritize them, large enough to be worth appealing.”
The panel
Market Reality Check AppealIQ already exists and operates on your exact model—contingency-based AI analysis of property assessments with filing automation. They're live, funded, and serving the same $500K–$5M segment. The live data shows 74% of property owners worry about taxes but only 22% have ever appealed, suggesting adoption friction is real but addressable. OverAssessed operates in five states (TX, GA, WA, AZ, CO), indicating geographic fragmentation is possible. Your core vulnerability: AppealIQ has solved the AI + contingency + filing playbook. Differentiation requires either superior hearing representation (human capital intensive, not scalable), tighter unit economics (hard when both charge ~20%), or geographic dominance in underserved counties. The market isn't saturated—most property owners still don't appeal—but the playbook is proven, meaning you're competing on execution and distribution, not invention. Bootstrapping works if you can acquire customers cheaper than AppealIQ; that's your real moat question, not the product.
Your underestimated technical challenge: comparable sales data is fragmented across county recorders, MLS systems, and proprietary databases with wildly inconsistent schemas and access restrictions. You'll spend 18 months building reliable data pipelines before your AI can even function. Counties don't standardize assessments—you're essentially reverse-engineering 3,000+ different systems. Build-vs-buy bite: Don't build your own hearing prediction model. You'll be tempted. Buy or partner with existing real estate valuation APIs (CoreLogic, CoStar, Zillow), then layer your contingency logic on top. Building from scratch means your model will systematically miss county-specific precedent that determines outcomes. No moat here. Once you prove the model works, any title company, tax service, or existing appeal firm replicates this in weeks. Your defensibility is pure execution speed and customer acquisition—not technology. What's well-chosen: the contingency model itself. It genuinely aligns incentives and solves the real friction point. That's your actual advantage, not the AI. The AI just needs to work well enough to be faster than a human paralegal reviewing comparables.
Your contingency model inverts the unit economics problem but doesn't solve it. CAC is brutal: you need to reach millions of fragmented property owners who don't know they're overpaying. Digital acquisition will cost $500–$2K per qualified lead; phone/local sales higher. LTV depends entirely on win rate and average savings. If you win 40% of cases and average $15K in first-year savings, you net $3K per win—meaning you need sub-$1,500 CAC to break even, which is unrealistic for cold outreach to this audience. The pricing assumption that's wrong: you're anchoring to traditional firms' 25–40% but haven't validated that property owners will accept 20% on contingency. They may prefer paying $2–5K upfront to a consultant who guarantees nothing but gives them optionality. Your contingency model shifts risk to you, not them—they have zero downside, which paradoxically makes you less trustworthy. Without traction, you'll burn 18–24 months proving win rates, average appeal values, and hearing success before a single dollar flows in. Runway evaporates fast. One thing working: the TAM is real and fragmented. Traditional firms' indifference to your segment is genuine white space—but it's white space because unit economics don't work at that property size without automation or a different pricing model.
Timing verdict: Late, but with a narrow viable window. Property tax appeals have been systematized for a decade—competitors like Reonomy, Assessor, and regional firms already serve this exact segment. You're not entering a nascent market; you're entering a crowded one where the unit economics (20% vs. 25-40%) are your only differentiator. That's a race to commoditization, not a timing advantage. Macro trend that matters most: AI-driven regulatory compliance automation. Counties are digitizing assessment records and automating hearing processes. This reduces friction for appeals but also narrows margins as the work becomes more standardized and less defensible. Window status: Closing. Existing players are already AI-augmented. Your contingency model doesn't solve the real friction—property owners still need to trust you, and trust scales slowly. Early-mover advantage is gone; you'd be a fast-follower in a maturing category. One genuine timing favor: Post-pandemic property value volatility. 2024-2026 saw significant market corrections in commercial real estate. Assessments lag reality by 1-3 years, creating legitimate appeal grounds right now that didn't exist in 2022. This window closes as assessments catch up.
Competitors found during analysis
Live dataAppealIQ
AI-powered appeals, contingency model, live
OverAssessed
Tax protest experts, 5-state coverage
Cause of death
AppealIQ Already Exists and You Have No Wedge
This isn't a "competitors exist" problem — it's a "your exact thesis has been implemented" problem. AppealIQ runs AI-driven contingency-based appeals targeting your segment. OverAssessed operates across five states. You're not bringing a new insight to market; you're bringing the same insight 18+ months later with zero traction, zero data, and zero county relationships. The question isn't whether competition is fatal — it's whether you have any differentiated angle at all. Right now, the answer is no. Same model, same pricing band, same segment, same technology stack. You'd need to articulate a specific wedge — a geographic stronghold, a distribution channel, a data advantage — and you haven't.
The Data Pipeline Problem Will Eat Your First Two Years
Your AI needs comparable sales data from county recorders, MLS systems, and proprietary databases across 3,000+ counties with wildly inconsistent schemas and access restrictions. This isn't a "scrape it and clean it" problem — it's an 18-month infrastructure build before your product even functions. Meanwhile, AppealIQ and others have already built these pipelines. You're not competing on AI sophistication; you're competing on plumbing. And plumbing that already exists elsewhere isn't a moat — it's a toll you pay while incumbents compound their data advantage.
Unit Economics Are Underwater at Your Target Property Size
Let's do the math the panel laid out. At a 40% win rate on $15K average first-year savings, you net $3K per successful case — meaning $1,200 per filed appeal on a blended basis. Customer acquisition for fragmented commercial property owners who don't know they're overpaying runs $500–$2K per qualified lead through digital channels. After you account for data costs, filing labor, and human hearing attendance, you're operating at break-even or negative margin on every case until you find a distribution channel that costs dramatically less than cold outreach. The traditional firms don't serve this segment for a reason: the economics are brutal without massive automation and cheap acquisition. You have neither.
⚠ Blind spot
You're assuming the contingency model is a trust accelerator. It's actually a trust obstacle at the point of sale. When someone offers you a risk-free service — "we only get paid if you win!" — the natural consumer response isn't gratitude, it's suspicion. What's the catch? Are you filing junk appeals to generate volume? Will you settle for a tiny reduction to collect your fee? The contingency model works beautifully after someone trusts you, but it creates friction before they do. Your real competitor isn't AppealIQ — it's inertia combined with skepticism. The 78% of property owners who've never appealed aren't sitting around waiting for a cheaper option. They don't appeal because they don't believe it works, and a stranger promising free money reinforces that skepticism rather than resolving it.
What would need to be true
You can acquire commercial property owners (or their managers) for under $800 per converted case — which almost certainly requires a channel partnership model, not direct digital marketing to fragmented owners.
Your win rate exceeds 35% in your target segment — validated across at least three states with different assessment methodologies before you invest in scaling the AI pipeline.
Assessment lag from the 2024–2025 commercial real estate correction persists through at least mid-2027 — giving you a window where the "properties are over-assessed" thesis is structurally true at higher-than-normal rates, buying time to build data infrastructure and prove unit economics before the window closes.
Recommended intervention
Stop trying to be a direct-to-owner platform. Partner with commercial property management companies — specifically firms managing 50–500 properties in the $500K–$5M range. There are roughly 30,000 such firms in the US. They already have the trust relationship with property owners, they're incentivized to reduce operating costs (it's literally their value proposition), and they can deliver you hundreds of properties per partnership instead of one at a time. Offer the management company a 5% referral cut from your 20% contingency fee, making it a profit center for them at zero effort. This solves your CAC problem (B2B2B instead of B2C), your trust problem (the management company vouches for you), and your volume problem (batch filings across a portfolio). Pick three states where assessment lag is worst right now — likely Texas, Illinois, and Georgia based on known assessment cycles — and sign five property management firms in each before you write a single line of AI code. Validate win rates manually first.
Intervention unlocking
5seconds
No account needed. One email, no follow-ups.
Want your idea examined? Free triage or full panel →