Oliver Alexander, CEO, Prophetic Software

98.5% of Land Is Never Listed

Oliver Alexander's Prophetic AI cuts land acquisition analysis from hours to six minutes for any parcel in the US.

 

AI-powered land discovery for real estate developers has moved from experiment to production. Prophetic Software - founded by Oliver Alexander in Portland, Oregon - enables development firms to search, filter, and analyze off-market parcels across all 160 million pieces of US land in a single platform, compressing analysis that previously took hours or days per parcel down to minutes. The platform combines AI-interpreted zoning data, demographic overlays, environmental screens, competitive pipeline mapping, and a land relationship management (LRM) system in one interface, replacing a workflow that most firms still run on county websites and spreadsheets.

 

Key Takeaways

• Only 1.5% of US land is listed on-market at any time. The other 98.5% - roughly 158 million parcels - is what Prophetic is built to surface. Every development firm competing for the same thin slice of listed inventory is competing against every other firm with the same data.

• Parcel analysis that takes trained analysts three hours takes six minutes on Prophetic. Alexander's own test: new hires - including PhD-level engineers - average three hours to answer ten basic questions about three randomly assigned parcels. On the platform, the same task takes six minutes.

• Zoning data exists in 9,000-plus named variants across US municipalities. Pre-AI technology could not solve this because keyword search fails when the same land use is described differently in every jurisdiction. Zone AI normalizes all of it into a single-click filter, with citations back to the source document.

• Handwritten outreach letters return a 12% response rate versus 0.5-1% for standard mailers. A Prophetic case study found response rates rise roughly twelve-fold when letters are sent via robotic-arm handwriting rather than printed mail.

• Site plan generation drops from one to five weeks of engineering time to two to five minutes. The Site AI tool generates a preliminary site plan - including open space, road placement, stormwater ponds, and environmental exclusions - sufficient to evaluate unit count before committing to engineering spend.

• Prophetic reached profitability after raising $3 million in venture capital. The company bootstrapped for its first 18 months before taking outside capital, then hit profitability shortly after. The business is currently hiring senior and full-stack engineers.

• The platform supports any asset class, not only residential. Clients include industrial outdoor storage developers in California, multifamily developers, and brokers covering a handful of counties. A custom cell tower filter was built for a single client request.

 

Adam Gower has spent more than 30 years working through exactly the kind of information asymmetries Prophetic is designed to remove. With $1.5 billion in transactional experience and clients collectively managing $45 billion in AUM, he has seen what it costs when development teams brute-force site selection through county databases and analyst hours. The question this episode addresses is straightforward: at what point does off-market land discovery become an operational necessity rather than a competitive advantage?

 

Why 98.5% of Land Stays Invisible to Most Development Firms

The US land market contains approximately 160 million parcels. At any given time, roughly 1.5% are listed on open-market channels. The remaining 98.5% - off-market, owner-held, and largely unanalyzed - is where Prophetic operates.

Alexander describes the current state plainly: large national homebuilders brute-force this problem by employing teams of analysts who work county websites parcel by parcel. Smaller developers and brokers lack both the staff and the infrastructure to replicate that approach. The result is that most of the development market competes over the same thin slice of available inventory.

"There's about 160 million pieces of land in the US. Then you look at how many of those are on market at any given time. That figure is about 1.5% of all those parcels. So everyone involved in real estate development, home building, all this stuff is kind of living and dying and fighting over the same crumbs."

County websites are part of the structural problem, not a solution to it. They are designed for individual parcel retrieval - a homeowner checking tax records - not for development firms running criteria-based searches across an entire metropolitan area. Firms trying to use them for off-market prospecting are forcing a public utility to perform a function it was never built for.

The legacy software market has not filled this gap credibly. Alexander notes that the dominant platform in the sector - an industry standard for two to three decades - was acquired by private equity and stopped updating, driving users back to county websites as the default. The Utility Thesis explains how companies like Prophetic are creating next generation AI industries.

 

How Prophetic's Zone AI Solves the Zoning Fragmentation Problem

 

US zoning is administered at the municipal level. Every municipality - save for a handful of exceptions, including Houston and certain unincorporated counties - maintains its own zoning code. Each code runs between 200 and 2,000-plus pages and uses its own terminology, format, and naming conventions.

The consequence for any technology trying to normalize zoning data across the country is that there is no shared vocabulary. Alexander put a number to the scale of the problem during a presentation at the Urban Land Institute in Raleigh: single-family detached housing alone can be described in over 9,000 distinct ways across US jurisdictions.

"That's why normal technology or pre-AI has never been able to solve this. Because you can't just look for single family detached or residential. It has to understand the context of the other descriptors and then bucket it as in or out of that category."

Zone AI addresses this through a multi-agent approach that extracts values from zoning manuals, interprets them in context, checks its own output, and cites the source document with a page number. Alexander is direct about why this is important: a developer who relies on an AI-generated zoning determination and commits capital based on an incorrect result has a serious problem. The cite-and-verify architecture is the engineering response to that risk.

The practical output is that a developer clicking any parcel in the country receives: the zone classification, the applicable land uses (permitted and conditional), minimum lot area with exceptions, setbacks, maximum height and coverage, ADU requirements, and a direct link to the relevant section of the municipal code. Tasks that would previously require an engineer or a planning attorney are answered on screen in a single click. This is where durable advantage in commercial real estate begins.

 

From Parcel Search to Closed Purchase: The LRM Workflow

 

Finding and analyzing a parcel solves one problem. Converting that discovery into a closed transaction requires a different set of capabilities. Prophetic's LRM - land relationship management - is the firm's answer to the second half of that workflow.

The system was built in response to client demand rather than planned product roadmap. Customers began telling Alexander's team that they needed the discovery and analysis tools to connect to their deal-tracking process. The LRM now integrates project management, contact tracking, task assignment, document storage, and outbound owner contact into the same interface.

The outbound piece includes a handwritten letter service that routes through a fulfilment partner using robotic arms to write physical cards in human-font ink. Alexander describes a case study with Polish Homes - a Pacific Northwest builder doing 500 to 600 homes per year - in which direct mail response rates went from roughly 0.5% to 1% for printed letters to approximately 12% for the handwritten format.

"When you get a custom handwritten card, I'm at least opening that and I'm going to see what's in there."

The rationale is simple. An off-market outreach letter implies intimate knowledge of the parcel. If the owner calls back and the development firm cannot immediately retrieve the project file, the credibility of the outreach collapses. The LRM closes that gap by keeping all parcel intelligence, contact history, and deal status in one place - retrievable in a single click from the project’s dashboard.

Alexander describes the current state at many large development firms as a recognition problem: "Your major multibillion dollar public divisions are doing this in Excel and they're doing it in their head." The LRM is positioned as the operational layer that connects discovery to execution. All this contributes to the productivity premium AI now delivers to sponsors.

 

Dev Map and Site AI: Market Intelligence Before Capital Commitment

Two platform components address the question developers face after they find a viable site: not whether they can build, but whether they should.

Dev Map is what Alexander describes as the first complete map of every new residential subdivision in the United States - spanning small towns, large cities, incorporated and unincorporated counties. Each project is clickable, displaying unit count, lot composition, MLS inventory status, closing velocity, and median sale price. For a developer evaluating a site, the tool answers the competitive context question: what is already being built nearby, how is it performing, and what pipeline is coming.

Alexander gives a direct example of how it is used in practice: "If I build 100 units and I didn't realize they're already progressing with 100 across the street, both of us putting 200 units online at the same time, we're both going to have a bad day."

Site AI addresses the unit count estimation problem. Before committing to engineering fees - which typically require one to five weeks of back-and-forth with an engineering firm to produce a preliminary site plan - a developer using Site AI can generate an estimate in two to five minutes. The tool inputs environmental constraints (wetlands, flood zones, topo), road entry points, stormwater pond placement, open space requirements, and density parameters, then outputs a preliminary site plan with lot count. Alexander demonstrated a 245-acre parcel that returned 415 lots in under five minutes.

The financial logic is direct: preliminary site plans are expensive to produce and are often used to determine whether a site is worth pursuing at all. Moving that decision point earlier in the process - before engineering spend - changes the economics of how many sites a development team can evaluate per month. 

 

Frequently Asked Questions

What types of real estate developers use Prophetic?

Prophetic serves any organization involved in ground-up development or land acquisition, including residential subdivision builders, multifamily developers, industrial outdoor storage operators, and real estate brokers. Alexander cites clients ranging from a two-county broker operation to national homebuilders doing tens of thousands of homes annually. The platform also has non-developer clients such as economic development agencies. Several asset-class-specific features - including residential subdivision tools - are built only for those use cases, but the core search, zoning, and LRM capabilities apply across all development types.

How does Prophetic handle AI hallucination risk in zoning data?

Zone AI uses a multi-agent verification process: one agent extracts the relevant values from the zoning manual, a second checks the output, and the system then cites the specific page number in the source document so developers can verify the result directly. Alexander acknowledges the stakes explicitly - an incorrect zoning determination can send a developer to the closing table on a site that does not support the intended use. The citation layer is the engineering response to that risk. It does not eliminate the need for professional zoning review before closing, but it replaces the hours of manual research currently required to produce a preliminary read.

What does Prophetic cost and how is pricing structured?

Prophetic prices geographically - by the coverage area a client needs, not by user count. A firm operating in one metro pays for that geography; a firm covering six states pays for that scope. Within a geography, all users sharing the same email domain have access to the platform under three tiers: Pro (entry-level), Expert (mid-tier), and Enterprise (white-glove onboarding and account management). Alexander did not disclose specific price points in this conversation. Interested parties can book a demo directly at prophetic-software.ai.

How long does onboarding take for a development firm?

Alexander reports that national homebuilders have been fully deployed in under one month. Practical proficiency for individual users takes approximately 90 minutes to two hours of training. The onboarding process includes importing existing projects and customizing the platform for the firm's specific asset class and geography. A follow-up session roughly one week later covers questions that emerge from initial use. After that, Alexander describes clients as self-sufficient. Pricing is structured to match the operational scope of each client, so smaller regional operators access the same tools as national builders within their geography tier.

 

Next Steps

Tier 1 - If you want to understand how AI is changing the operational side of CRE before committing to any platform, start by subscribing to the free GowerCrowd newsletter.

Tier 2 - If your firm is actively evaluating AI systems for deal sourcing, site analysis, or capital deployment, the AI for Real Estate Accelerator executive program is where that work gets done with practitioners, not theorists.

 

Adam Gower, Ph.D. is the founder of GowerCrowd and one of the most experienced practitioners in commercial real estate capital formation. With more than 30 years and $1.5 billion in transactional experience - including serving as President of a Universal Studios development division overseeing $400 million in projects across Asia Pacific - Adam now helps CRE sponsors build AI-powered systems for investor acquisition, deal management, and capital deployment. His clients collectively manage over $45 billion in assets under management. gowercrowd.com