Moat & Differentiation¶

The honest question: if Housing.com decided to copy PropPie tomorrow, what stops them?

This doc is the inventory of what stops them — and what doesn't.

The five layers of the moat (in order of durability)¶

flowchart TD
    L1[Layer 1: Business-model alignment<br/>buyer-paid, not seller-paid] --> L2
    L2[Layer 2: Compliance posture<br/>info-only by design] --> L3
    L3[Layer 3: Data depth<br/>Maharashtra fusion, attribute-level] --> L4
    L4[Layer 4: Citation + provenance discipline<br/>every claim traceable] --> L5
    L5[Layer 5: AI-native UX<br/>conversational, probabilistic, simulations]

Each layer requires the one below to be credible. You can copy L5 in 6 months. You cannot copy L1 without restructuring your company.

Layer 1 — Business-model alignment (MOST durable)¶

The moat: PropPie is paid by users (subscriptions, listing fees from issuers we curate, B2B SaaS) — not by developers for placement.

Why it's durable: Housing.com, MagicBricks, NoBroker, Square Yards all derive 70-90% of revenue from developers paying for listings, leads, and ads. To copy our positioning, they'd need to: 1. Tell their developer customers they're not the primary customer anymore 2. Replace ~80% of their revenue with subscriptions and B2B 3. Retrain a sales org that has decades of muscle memory on developer relationships

The economics simply don't work for them. A new entrant could do it (someone might), but no incumbent will. This is structural.

Risk to the moat: Pressure from investors to "monetise faster" via developer ads. We don't do this. Locked into the SOUL document.

Layer 2 — Compliance posture (very durable)¶

The moat: PropPie operates as an information service — not an Investment Adviser, not a Research Analyst, not a real-estate agent (yet). Every product surface is designed to live cleanly inside this posture.

Why it's durable: Once an incumbent has been operating in soft "you should buy" territory, restructuring to info-only is harder than starting there. Existing chatbots from Housing.com casually slip into recommendation language. To remove that is a brand promise downgrade. To not remove it is an SEBI risk.

Meanwhile, we never need to register, never need a compliance officer for IA, never need NISM-X qualifications for personnel, never need to disclose AUA. Our compliance posture enables speed.

Risk to the moat: SEBI issues guidance that pulls more conversational tools into IA territory. We'd then need to register OR re-trim our outputs. Both are possible; the latter is faster.

How to deepen this moat: publish (informally) our compliance principles, get a public legal opinion, position openly as "the proptech you can trust because we built compliance in."

Layer 3 — Data depth (durable, requires investment)¶

The moat: The fused Maharashtra data layer — MahaRERA project files, IGR registered transactions, MahaBhulekh land records, Government Resolutions, MRSAC GIS, news + social — with entity resolution, conflict resolution, lineage, and confidence per attribute.

Why it's durable: This is years of engineering work. The data is public but messy. Building the fusion right takes:

Robust scrapers for MahaRERA across project-page format generations
LLM-assisted extraction from scanned PDFs in mixed Marathi/English
IGR ingestion at sub-day latency for priority micromarkets
Entity resolution: linking a promoter's PAN across 14 different project filings + their MCA records + their court cases
GR classification: discriminating real-estate-relevant GRs from the 80% noise
Quality framework: confidence, lineage, freshness SLA per attribute
~140 derived attributes with explainable, deterministic computation

Most engineering teams underestimate the depth. Vishal's pipeline is the spine.

Risk to the moat: PropEquity has 15 years of data we don't have. They can press their advantage. We win on depth-per-attribute in Maharashtra, not breadth-of-states.

Layer 4 — Citation and provenance discipline (durable)¶

The moat: Every claim in user-facing output links back to its source. Every derived score has visible inputs and weights. RAG with deterministic provenance.

Why it's durable: It's a discipline as much as a technical investment, but technically: implementing RAG that always cites correctly, with the document image / PDF page rendered inline, is non-trivial. Most LLM products lie about sources or omit them.

This is also a brand moat. Once "PropPie shows its work" is the user expectation, an unsourced incumbent looks shifty by comparison.

Risk to the moat: Frontier LLMs get cited-RAG natively (already happening) → tech moat erodes; brand becomes everything. Plan for that.

Layer 5 — AI-native UX (medium-term)¶

The moat: Conversational interface for B2C; simulation + counterfactual workflows for B2B; probabilistic projections; "explain like I'm 5" toggles; multilingual.

Why it's not super-durable: Anyone with frontier-model access can build a chatbot. The UX patterns will get copied. The wow-moments catalog is a moat for ~12-18 months.

What to do: Use the early lead to lock in brand and habit. Don't bet the strategy on this layer.

What is NOT a moat (despite feeling like one)¶

"AI" itself. It's table stakes now. Saying "we have AI" in 2026 is like saying "we have a website" in 2010.
Pretty UX. Will be copied.
The brand name "PropPie". Memorable, but trademarks are not strategy.
Founder relationships. Helpful, not a moat.
A specific LLM choice. Vendor lock-in is the opposite of moat.

Sequencing the moat¶

Foundation phase (now → Q4 2026): - Build L3 (data depth) hard — this is the long pole - Lock in L1 and L2 in every product decision — SOUL + compliance discipline

Growth phase (Q1 2027 → 2028): - Compound L4 with discipline; make citation a brand signature - Use L5 to win launch attention, then transition to brand-driven growth - Make L1 visible in marketing: "Why we don't take developer commissions"

Defense phase (2028+): - Deepen L3 to other states (Karnataka, Gujarat) — defensive against pan-India entrants - L1 + L2 become institutional — board-level commitments

How a competitor would attack us, and how we defend¶

Attack vector	Likely attacker	Defence
"We're also source-cited now"	Housing.com adds citation badges	We're cited at the attribute level with confidence; theirs is shallow. Show the depth.
"We have AI too"	NoBroker / Housing chatbots	Theirs leak into recommendation; ours never does. Position the discipline.
"We have more pan-India coverage"	PropEquity, CRE Matrix	True at breadth; we win on depth + freshness + UX in Maharashtra. Don't dilute.
"Cheaper subscription"	New entrant	Price-only competition we can match; quality + reliability + citation become the differentiation
Acquihire / aqui-poach	Any	Talent retention through ownership, mission, and SOUL alignment
Regulatory shadow	A bad-actor competitor's incident triggers blanket scrutiny	Our clean posture is the defence; document compliance publicly

What we should publish / signal externally¶

To make the moat legible (and self-fulfilling):

A "trust report" twice a year: how many citations, how many corrections, how many flagged transactions
The SOUL document, lightly edited, on the website
Compliance posture page (we don't sell user data, don't take seller commissions, etc.)
"Why we won't tell you what to buy" — explainer post
Open-source the Maharashtra data dictionary (the structure, not the data) — community trust signal

The one-sentence moat¶

"We're the only proptech in India whose business model, compliance posture, data depth, citation discipline, and AI-native UX are all aligned to serve the buyer — and any incumbent that copies us has to break their business model to do it."