NakedHouse.ai reveals what a property can legally become
Its embedded development rights under current zoning, packaged as a per-address Good News Report with visual concept plans, delivered where property decisions already happen: inside the real estate transaction.
Zillow shows what a property is. NakedHouse.ai shows what it can become. No listing portal offers a per-address buildable envelope, floor area ratio, or setback determination. That gap is the wedge.
One line per segment
Listing agents
Walk into every listing presentation with the one report no competing agent has, attached to the NHD order they already place.
Homebuyers
Before you buy, know what the lot can become: the upside Zillow cannot show, for $99 instead of a $2,000 architect study.
Homeowners
You may be sitting on unclaimed buildable rights. Find out what your property is legally entitled to, in minutes.
Investors and developers
Parcel-level envelope and multi-scenario analysis at $299 to $499, without the NYC-only footprint or the $8,000 per year toolchain.
Lenders and insurers
Address-level development-rights intelligence as the upside twin of property-risk data. A year-two conversation, not a launch segment.
The three moves that matter
- Attach the report to the NHD order. Property I.D. already sells a required report on essentially every California residential sale. The Good News Report rides the same order, payment, and delivery pipes at a comparable price, so incremental acquisition cost is near zero. The first 90 days exist to measure one number: the attach rate.
- Activate only the warm base, with humans in the loop. Outreach goes exclusively to the productive top quartile of existing NHD-ordering agents. AI drafts from each agent's own listings; a human approves every send. No cold outreach and no autonomous sending in year one.
- Build the answer-engine surface, not paid search. Top-of-funnel money goes to being the cited answer for "what can I build on my lot" in ChatGPT, Gemini, and AI Overviews. City and legislation explainers are published openly; the per-address value stays behind the $99 report.
The growth flywheel
Each turn of the loop lowers blended acquisition cost and funds the next turn. The first link inherits transaction-volume cyclicality; homeowner demand and AI citations are the two spokes designed to spin even when transactions do not.
The Good News Report is an informational report, not a permit or entitlement guarantee, and it never implies economic feasibility.
The operating model
A small number of humans making judgment calls on top of an AI agent fleet doing the volume work. The fleet researches, drafts, triages, scores, and answers. The humans approve, sell, certify, and escalate. Nothing customer-facing ships without a human decision where it matters.
At steady state, months 9 to 12, the fleet is roughly 50 agents in six pods.
The human layer
Always human, no exceptions
- Final send approval on any outbound message.
- All pricing exceptions and discounts.
- Any claim about what can legally be built.
- Any escalated support call.
The fleet drafts and queues. Humans decide.
Phase 1: Days 0 to 30, prove the attach on live transactions
The only number that matters this quarter is whether a discretionary $99 report attaches to a required NHD order. Everything in this phase exists to buy that number cheaply.
- Integrate the one-click add-on. When an agent orders an NHD, the confirmation screen and email show a pre-computed upside teaser for that address and one button: add the Good News Report for $99. Same order, same payment rail, same delivery. Engineering owns weeks 1 and 2; legal reviews carrying a non-disclosure product as an escrow line item in parallel.
- Pick the pilot cohort. Rank the Property I.D. agent base by trailing 12-month NHD order volume. Select top-quartile producers in 2 to 3 California metros with strong ADU volume, for example Los Angeles and San Diego. Target 25 to 50 named agents. These are warm customers, not prospects.
- Build the report pipeline with the human QA gate. Licensed parcel and zoning data in, envelope and floor-area analysis, concept plan rendering, then the planner or architect reviews every claim before anything ships. Cost this pipeline honestly in week 2; per-report cost including QA is a kill threshold later.
- Stand up the first two agent pods. Research, for data QA on the pilot metros so the teaser flag is right before it is ever shown, and content, for the report-page QA workflow and the first co-branded one-pager template.
- Instrument everything. Offer shown, offer clicked, report purchased, report delivered, agent reorder. Attach rate is paid reports divided by NHD orders where the offer was actually presented. Build the dashboard before the first offer goes live.
- Ship co-branded samples to pilot agents. Each agent gets a free sample report on one of their own active or recent listings, with their photo and brokerage on it, plus the one-pager. The sample is the pitch.
- Run the weekly cadence. Every Monday: attach numbers by agent and metro, QA error log review with every catch root-caused, pipeline cost per report, and one decision about the offer. One iteration per week, no more.
Expected results after 30 days
Pilot live with 25 to 50 top-quartile agents in 2 to 3 metros. Add-on live in the order flow. First paid reports delivered, target 30 or more. An early attach reading on a small but real base. Per-report cost measured. QA error rate of zero on shipped reports.
Phase 2: Days 31 to 90, build the outbound and voice motion
Phase 1 proved the mechanic can exist. Phase 2 builds the muscle that scales it, still entirely inside the warm base.
- Turn on the SDR pods, warm base only. AI drafts, human approves, roughly 20 to 30 approved sends per rep per day, each personalized from the agent's own listings run through the teaser engine. No cold outreach to non-customer agents in year one. No unverifiable buildability claims in any copy.
- Launch the inbound voice agent. It answers report questions, gives order status, and qualifies inbound leads. It escalates anything it cannot resolve in two turns, anything legal, and any upset caller. Every escalation is recorded and reviewed weekly to tune the boundary.
- Bring CRM agents live. They score every reply and teaser user, keep records clean, and queue a ranked daily call list so the human reps start each day with priorities, not a blank CRM.
- Run brokerage-team pitches. Human reps pitch the report as a listing-winning tool in a post-settlement market where presentations are contested. One team deal converts one warm relationship into ten. Bundle packs, for example 10 reports at a pack price, are the close.
- Put the free teaser on the web. Enter an address, get a coarse, banded upside flag: upside exists, estimated buildable class, nothing actionable. The teaser must never answer "what exactly can I build." Keep the band coarse and watermark outputs.
- Publish the first answer-page wave. 10 to 15 city and legislation explainers timed to the SB 79 effective date and the 2025 CEQA reforms, with direct answers up top, named methodology, and dated sources. Measure citation share weekly against a fixed query basket. Zero paid search.
- Iterate the offer twice. A/B the bundle discount depth and the $79, $99, and $149 price points inside the pilot. Two honest iterations before any pivot conversation.
Expected results after 90 days
Attach rate at or above 5 percent of offered NHD orders, with 8 to 10 percent validating the model. Measured reply and meeting rates from warm outbound. First brokerage team deals signed. Teaser live with a conversion reading against the 5 to 15 percent sales-assisted band. Voice agent resolving a measured share of inbound without escalation. Baseline citations on at least 5 queries in the basket.
Phase 3: Months 4 to 6, scale what worked
The 90-day readout tells us which pods and which motions earn their keep. This phase spends on winners and kills losers.
- Expand metros. Add the next 2 to 3 California metros by ADU volume and top-quartile agent density. The activated cohort grows from 25 to 50 agents toward 200 or more.
- Scale winning pods, shut the losers. Any pod that cannot beat its human-baseline cost per outcome within 60 days of launch gets rebuilt once or killed. Cost per approved send, per booked meeting, per resolved call, per shipped page. The operator owns this scorecard.
- Launch the investor tier. $299 to $499 on the same pipeline: multi-scenario analysis, same data, same QA gate, deeper output. No second product and no separate go-to-market; investor-behaving users get routed to it by the CRM agents.
- Build the builder and architect referral loop. Every delivered report offers a warm handoff to 2 to 3 vetted local builders or architects, with a referral fee back to NakedHouse. The report is honest about being informational; the referral gives "what next" a real answer.
- Run the lender pilot. 2 to 3 conversations with construction and HELOC lenders: a homeowner with documented buildable upside is an origination prospect. Pilot only, no dedicated headcount, no dependence.
- Start the homeowner-direct hedge. Small, controlled paid tests in ADU-heavy zip codes, fed by the teaser and the answer pages. This is the deliberate second segment, funded because the agent rail proved its economics, not instead of proving them.
Expected results by month 6
Attach rate at or above 8 percent sustained on the expanded cohort. 200 or more activated agents. 500 or more paid reports per month. Investor tier live at around 10 percent of revenue. First referral fees collected. Homeowner-direct at roughly 10 percent of report volume. Per-report cost including QA holding under the 30 to 40 dollar threshold. At least one pod killed or rebuilt; if nothing has been killed, the bar is too low.
Phase 4: Months 7 to 12, compound
- Expand out of state through title and escrow partners. Take the proven California attach data to the entities that occupy the transaction moment in Texas, Florida, and Washington. The partner's first question is "what does it attach at," and we now have the answer. Sign the first partner; never let a partner sit between NakedHouse and the interpretation and QA layer.
- Defend the answer-engine position. Measure citation share weekly, target being cited on half the query basket, publish tranches 2 and 3 on legislation effective dates, and monitor free chatbots monthly for teaser reconstruction. Coarsen the band before ever touching price.
- Bring the fleet to full 50-agent steady state. The operator's scorecard governs the mix; if follow-up agents outperform SDR agents per dollar, the ratio shifts.
- Tune the customer-service blend. Voice agents resolve the routine half of inbound: order status, delivery questions, basic report walkthroughs. Humans take the rest, with escalation review keeping the boundary honest.
- Package the year-one dataset. Attach rates by metro, teaser conversion, referral economics. This is the pitch asset for year-two partnerships and the lender and insurer conversations.
Expected results by month 12
Attach rate at or above 10 percent across the activated top-quartile base. 500 or more activated agents. 1,500 or more paid reports per month. Investor tier at 15 to 20 percent of revenue. Homeowner-direct at roughly 20 percent of volume. First out-of-state partner signed. AI referral a top-3 direct-channel source, small in absolute volume but compounding. Shipped QA errors still at zero.
The stack
| Component | Buy or build | Call |
|---|---|---|
| CRM | Buy | HubSpot, already in-house at Amberoon. The CRM agents work through its API rather than a new system. |
| Outreach tooling | Buy the plumbing, build the brains | A standard sequencing tool for send and deliverability; the personalization drafting is our own agents on our own listing data. |
| Voice agent platform | Buy | A commercial voice AI platform of the Retell or Vapi class with human warm transfer. Building telephony is not our business. |
| Parcel and zoning data | Buy | License Regrid-class standardized zoning and parcel data. Raw data is a commodity; do not collect it ourselves. |
| Report rendering pipeline | Build | This is the product: envelope analysis, concept plans, QA workflow. Own every line of it. |
| Teaser engine | Build | Coarse upside flag computed from licensed data at near-zero marginal cost. Validate that cost assumption in the first build sprint. |
| Analytics | Buy | GA4 plus a product analytics tool, with one custom attach-rate dashboard on top. Do not build an analytics platform. |
Hiring
| Role | When | Owns |
|---|---|---|
| Fleet operator | Month 0 | Agent pods, guardrails, cost per outcome, kill decisions. First hire. |
| Planner or architect, QA lead | Month 0 to 1 | The human QA gate on every shipped claim. Non-negotiable before the first paid report. |
| Salesperson 1 | Month 1 | Pilot agent activation, then brokerage-team deals in metro 1. Ideally an existing Property I.D. agent-facing rep. |
| Salesperson 2 | Month 3 | Second metro group, bundle packs. |
| Customer success and service lead | Month 3 | Voice agent supervision, escalations, post-purchase experience. |
| Salesperson 3 | Month 6 | Only if the 90-day attach readout validates. Works expansion metros. |
| Second QA reviewer | Month 6 to 9 | Volume relief on the QA gate so throughput scales without loosening it. |
Five hires by month 3, seven by month 9. Everyone else is fleet.
What we measure
- Attach rate of offered NHD orders. The primary number. Floor is 5 percent by day 90; 8 to 10 percent validates the model; 10 percent or more across the activated base by month 12.
- Teaser-to-paid conversion. Against the 5 to 15 percent sales-assisted band in agent context; 3 to 5 percent expected on cold web traffic.
- Cost per meeting from warm outbound. Must run meaningfully below cold benchmarks. If warm outreach performs at cold rates, treat that as a relationship-quality finding and shift weight to brokerage-team deals.
- Report QA error rate. Target is zero shipped errors. Catches at the gate are logged and root-caused weekly. One shipped wrong buildability claim is a brand and liability event, not a metric blip.
- Revenue per transaction. Bundle plus report plus referral fee. The counterweight to a 30-year-low transaction market is growing revenue per transaction, not waiting for volume.
- Investor-tier mix. Around 10 percent of revenue by month 6, 15 to 20 percent by month 12.
Two kill triggers, decided in advance
Kill trigger 1: the rail thesis
Attach below 2 to 3 percent after two honest offer iterations means the rail thesis failed. Pivot the channel: shift weight to homeowner-direct, answer-engine content, and the investor tier.
Kill trigger 2: the margin thesis
Per-report production cost above 30 to 40 dollars with human QA at the $99 price means the margin thesis failed. Re-engineer the pipeline or move price up-band before scaling. Do not scale unprofitable reports.
Why the plan looks the way it does
Every step in the execution plan traces to one of six argued decisions. Here is each decision, the alternatives considered, the evidence, and the result we expect. Single-source and vendor-biased figures carry their flags.
Decision 1: Win listing agents first
Real intent exists: 82 percent of Americans use AI for housing insights, and 86 million homeowner households are not gated by transaction volume. But there is no distribution asset there yet, cold freemium converts at 2 to 5 percent, and the proptech graveyard is full of consumer-direct plays that died on acquisition cost, Reali among them with more than $290M raised.
Highest willingness to pay, with the category priced at $50 to $300 per parcel SINGLE-SOURCE and Tectmind at $499 or more per report. But credible substitutes exist at every price point, so entering there abandons the one unique asset.
Moody's acquiring Cape Analytics proves institutions pay enterprise money for address-level property intelligence, but those sales are long-cycle and opaque-priced. A place to exit to, not to start.
The channel is structurally decaying: NAR membership is below 1.5 million and falling, and nearly 3 of 4 agents closed zero transactions in 2024. That modifies targeting, to the productive top quartile, rather than reversing the choice.
Target the productive top quartile of Property I.D.'s existing NHD-ordering agents, with the homebuyer as the paying end-consumer. It is the only option where the sole non-replicable asset, the transaction rail, does the work: 67 percent of agents want better analytical AI while only about 22 to 39 percent have it.
Expected result: a pilot cohort that actually transacts, so every offered order is a real at-bat.
Sources: RPR, Realtor.com via Inman, Real Estate News, TechCrunch, Tectmind
Decision 2: Transaction-embedded cross-sell as the primary channel
Fast and testable, but no benchmark in the corpus supports paid economics for this product. What the corpus contains is the negative case: Reali's death featured high acquisition cost, and cold acquisition converts at 5 to 20 percent against 60 to 70 percent warm.
Partnerships genuinely scale, as Gridics and LandLogic SINGLE-SOURCE show, but they are slow, they insert a partner between NakedHouse and its margin, and Property I.D. already is the embedded partner other startups would have to sign. Doing partnerships first would be paying for what is already owned.
The rail is California-only and inherits the transaction trough: about 269,000 existing single-family sales per year, flat. That argues for a hedge, not a different primary.
Ride the NHD order flow. Selling to an existing relationship converts at 60 to 70 percent versus 5 to 20 percent cold, and the order, billing, and delivery pipes already exist, so incremental acquisition cost is near zero. Point-of-sale attach lift figures of 2 to 3 times exist but carry high vendor bias; plan on the defensible 5 to 20 percent band. The escrow line-item mechanics for a non-disclosure product are a flagged legal work item, not an assumption.
Expected result: an attach-rate answer bought for a fraction of what any paid channel would cost to test.
Sources: Marketing Metrics via Upland Kapost, C.A.R. 2026 Forecast, Extend, Gridics
Decision 3: Answer-engine presence ahead of paid search and classic SEO
The upside cases are vendor showcases, and the downside tail is documented and catastrophic: Google's March 2024 scaled-content update demoted template location pages en masse, HouseFresh fell about 95 percent, and 32 percent of 671 travel publishers lost more than 90 percent of traffic SINGLE-SOURCE. Open per-parcel pages would also give the product away to train Google.
No supporting economics anywhere in the corpus, and it competes on cost against a free structural channel and a free high-conversion channel.
AI referral is roughly 1 percent of traffic today. This channel seeds a compounding position in year one; it does not fill pipeline, and the transaction rail never depends on it.
Publish city and legislation explainers engineered to be the cited answer for "what can I build on my lot," and keep the per-address value gated. AI-referred visitors convert 2 to 11 times better than search traffic, 82 percent of Americans already use AI for housing insight, and the audience's research now starts inside the answer engine.
Expected result: a small but compounding channel seeded in year one that the transaction rail never depends on.
Sources: Microsoft Clarity via Digiday, DigitalApplied, SE Ranking, Bain, Realtor.com via Inman
Decision 4: AI-augmented human outreach, not autonomous AI SDRs
The one controlled study available, 100,000 paired emails, shows autonomous AI trailing humans on reply by 21 percent and meetings by 36 percent, with spam flags at 8 versus 3 percent and inbox placement at 71 versus 86 percent and worsening as filters adapt. The seemingly contrary 18 to 22 percent reply numbers describe AI-assisted humans, not autonomous AI. Strategically worse: spam cadence aimed at the warm base would spend the company's one irreplaceable asset to save SDR salary.
Humans win on every quality metric, but pure human does not scale to the productive quartile at a $99 average sale. Cost per qualified opportunity ran $487 human-only versus $224 hybrid in the one economics comparison available SINGLE-SOURCE.
Hybrid, warm-first. Every SDR agent drafts and a human approves. Hybrid pods cut cost per qualified opportunity roughly in half and booked 1.9 times more meetings per dollar than pure AI, figures that are vendor-sourced and used as direction, not plan-of-record math. The warm-base restriction converts the problem from the 5 to 20 percent cold regime to the 60 to 70 percent existing-relationship regime, a bigger lever than any SDR technology choice.
Expected result: warm-regime reply and meeting rates without spending the trust of the one audience the whole thesis depends on.
Sources: DigitalApplied, Salesmotion, Optifai
Decision 5: A non-substitutable teaser instead of freemium
Freemium converts at 2 to 5 percent, cross-verified across three sources, and every full report carries real marginal cost in data, rendering, and QA. The 95 to 98 percent of never-converting free users become a per-unit cash drain worse than SaaS.
The category is unknown to consumers, so with zero free surface every sale requires the agent to explain an unfamiliar product cold, and the answer-engine channel is starved of citeable surface.
Paid per report, fronted by a coarse, banded upside flag at near-zero marginal cost. The free layer must never answer "what exactly can I build," only "there is something here worth $99 to see." Known risks are flagged: the near-zero cost of the flag is an engineering assumption to validate in the first sprint, and free chatbots reconstructing the teaser is a watch item, answered by coarsening the band before touching price.
Expected result: an education and sharing surface at near-zero marginal cost that feeds the attach flow instead of cannibalizing it.
Sources: Userpilot, ChartMogul, First Page Sage
Decision 6: The $99 price
Solves a problem the evidence says does not exist: agents rank cost near the bottom of AI concerns while accuracy is first at 63 percent. Underpricing forfeits margin the low-volume market makes precious and signals commodity in a category where free chatbots already give away text answers.
The attach mechanic depends on an easy yes inside a transaction where the NHD anchor is roughly $70 to $100; a $499 line item breaks that logic. Tectmind's $499 is a pro product in the most expensive US zoning market, New York, the wrong reference class for a consumer attach.
$99 per Good News Report, defensible band $79 to $149, plus a $299 to $499 investor tier on the same pipeline. The price sits at NHD parity, standard around $69.95 and premium $99.95, while running 20 to 40 times under the architect-study substitute that defines the value story. Confidence is high on the band and medium on the exact point; no willingness-to-pay study exists in the corpus, so pilot tests stay inside the band. The $50 to $300 AI-zoning category band is directional SINGLE-SOURCE.
Expected result: an easy yes inside escrow that still funds human QA, with the investor tier capturing pro buyers on the same pipeline.
Overall assessment
The strategy concentrates on the single non-replicable asset, the NHD transaction rail, and aligns with cross-verified evidence at every decision point. But its central load-bearing number does not yet exist. Four unknowns determine the outcome:
- The discretionary attach rate on the NHD rail. The closest analog, home warranties at closing, has no published attach rate at all. At California's roughly 269,000 annual sales, a 10 percent attach at $99 is a real business; at 2 percent it is not.
- Effective channel size. About 75 percent of agents closed zero transactions in 2024, so the true universe is the productive quartile, and transaction volume sits at a 30-year low.
- Accuracy and liability discipline. One wrong buildability claim is a brand and liability event; human QA is a cost and a gate on scale, but non-negotiable.
- The demand thesis surviving feasibility reality. SB 9 produced 266 projects from roughly 6.1 million eligible parcels: legal permission is not economic feasibility. The ADU wave, about 1 in 5 new California homes, says the rights are real; SB 9 says overselling them destroys trust.
What upgrades the assessment: pilot attach at or above 8 to 10 percent on live NHD orders, or answer-engine citation share materializing with measurable sign-ups.
What downgrades it: pilot attach below 2 to 3 percent, per-report cost with QA above 30 to 40 dollars at the $99 price, a platform policy shift that stops citing gated commercial reports, or evidence that the NHD relationships are order-processing rather than trust relationships.
The field in three layers
Reference material. The competitive field splits into three layers, and NakedHouse's wedge does not sit squarely in any one of them.
Zoning data and AI chat tools
Zoneomics, ZoningChat, My Zoning AI, ZoningBot. Cheap or self-serve. They tell you the rules, not a polished consumer "here is what you can build" report with renderings.
Developer and CRE feasibility platforms
Gridics, TestFit, Deepblocks, Tectmind, CityBldr. Massing and financials for professionals doing deals, at professional prices. Not for homeowners or agents at the point of listing.
Substitutes
Zillow and Redfin show what a property is. Architects run $2,000 feasibility studies. City planning counters are authoritative but slow. NHD incumbents already ride the transaction NakedHouse attaches to.
Nobody owns the consumer development-rights report at the transaction point. The closest feature match, Tectmind, is New York only and pro-priced. The closest distribution match, the NHD incumbents, does not sell development upside. NakedHouse sits in the white space between them.
Comparison
| Player | Who it serves | Price | What you get | Coverage |
|---|---|---|---|---|
| NakedHouse.ai | Listing agents, buyers, homeowners; investor tier | $99 report; $299 to $499 investor tier | Per-address development-rights report with buildable envelope and visual concept plans, human QA on every claim | California at launch on licensed standardized zoning, national path via the same data |
| Zoneomics | Brokers, planners, appraisers, lenders | $61 to $186 per month; briefs $39 to $65 | Zoning rules, briefs, and a chat assistant; the rule set, not a consumer narrative with renderings | Self-reported 10,000+ cities, 100M+ parcels SINGLE-SOURCE |
| Gridics | Municipal governments; CRE users in calibrated cities | Opaque; municipal contracts and subscriptions | Code digitization for cities, parcel feasibility for pros in covered markets | City-by-city since 2017; freshest third-party coverage dates 2017 to 2018 |
| Tectmind | NYC developers, owners, investors | Reports from $499; expert review from $1,500 | The closest feature match: parcel zoning analysis with massing and financial view, plus consult | New York City only |
| TestFit and Deepblocks | Developers and CRE teams doing many deals | TestFit Site Solver from $8,000 per year; Deepblocks from $2,000 per month | Generative site planning and pro formas; overkill and overpriced for a one-off consumer question | Multi-market, pro-grade |
| Architect feasibility study | Homeowners and developers ready to spend | Around $2,000 flat, or $100 to $250 per hour | The direct human substitute: bespoke options, budget, and schedule over days to weeks | Local, relationship-driven |
| City planning counter | Anyone; the official path | Roughly $54 to $194 per letter or hourly meeting fees | Authoritative determination, but slow, appointment-driven, non-visual, jurisdiction by jurisdiction | One jurisdiction at a time |
| Zillow and Redfin | Mass-market buyers, sellers, owners | Free | What a property is: facts, photos, estimates, ADU explainers. No per-address buildable envelope, floor area ratio, or setback determination | National |
| Permit certainty: none of the above, NakedHouse included, guarantees a permit or entitlement. Every player in this table sells information or analysis; only the city issues approvals. NakedHouse states this on every report. | ||||
Sources: Zoneomics, Gridics, Tectmind, TestFit, Handel Homes, Monograph, Zillow, MyNHD
Next best option by segment
| Segment | If not NakedHouse, then |
|---|---|
| Listing agent | A Zoneomics brief at $39 to $65 for the rule set, minus the consumer narrative and renderings; or nothing, which is the current default. |
| Homebuyer | Zillow's ADU explainer plus a free chatbot answer, then the wall: no per-address determination without paying an architect. |
| Homeowner | The city planning counter or a $2,000 architect feasibility study, both slower and one of them 20 times the price. |
| Investor or developer | Tectmind at $499 in New York, ZoneIQ in Gridics-calibrated cities, or TestFit and Deepblocks subscriptions for high-volume work. |
| Lender or insurer | Cape Analytics-class risk intelligence, now inside Moody's; nobody sells them the upside twin yet. |
Where we are not differentiated
- Raw parcel and zoning data. Regrid sells standardized zoning across roughly 14,000 municipalities to anyone, us included. Data is a commodity input, not a moat.
- Free text answers. AI chatbots already answer zoning questions in plain language at zero price. We do not compete with free text; we compete on validated per-address interpretation, visuals, and QA.
- Permit certainty. Nobody in the field, us included, can promise an approval. Differentiation on honesty is available precisely because the claim cannot be made.
Sources: Regrid, yeschat, Insurance Journal