AI UX Implementation Plan¶
Combined view: Aishvarya's 12 UX ideas + our 24 wow moments. Mapped to available MahaRERA + IGR fields. MVP-ready items first.
Business Summary¶
PropPie's AI UX layer turns two government data sources (MahaRERA APIs + IGR Index-II) into investment intelligence that no Indian proptech offers today. The core differentiator: every number traces back to a registered deed or RERA filing — not broker opinion. Phase 1 delivers 10 implementable screens using only the data we already have. Phase 2 adds GIS, news/sentiment, and GR feeds.
The 18 Unified AI Ideas (deduplicated from Aishvarya's 12 + our 24)¶
Tier 1 — Buildable NOW from MahaRERA + IGR data alone¶
| # | AI Idea | Wow Factor (2 lines) | MVP from current fields? |
|---|---|---|---|
| 1 | Transaction Explorer + Source Trail (Aishvarya: Intelligence Dashboard + Our: A1) | Every price claim links to the actual IGR deed. Users can drill from a median to the 87 transactions that produced it — nobody else shows their math. | Yes — IGR sale_price, sale_date, parties, area, survey number |
| 2 | Investment Dossier (Aishvarya: Idea 2 + Our: A3, B1) | One-click PDF report per project: RERA details, cost breakdown, risk flags, comparable transactions, all source-cited. Replaces weeks of manual research. | Yes — MahaRERA project details + IGR transactions matched by survey number |
| 3 | Developer Trust Score (Our: A4, C5) | A transparent 0-100 score decomposed into delay, complaints, cost overruns — each bar links to the RERA filing. No black box. | Yes — MahaRERA: completion dates, complaints, extensions, litigation |
| 4 | True Cost Calculator (Our: B1) | Shows the real ₹99.3L cost of a "₹85L" flat — stamp duty, registration, GST, maintenance, parking. Jaw-drop moment. | Yes — IGR stamp_duty, registration_fee + deterministic rules per district |
| 5 | Carpet Area Reality Check (Our: B2) | Catches brochure inflation instantly: "brochure says 1,180 sqft but RERA filing says 718 sqft carpet." Exposes the industry's biggest trick. | Yes — MahaRERA area fields vs user input |
| 6 | ASR Gap Scanner (Aishvarya: implied in Risk Engine + Our: C4) | Flags when a sale price is 110% above government valuation vs the typical 60-80% — instant overpricing/distress signal from registered data. | Yes — IGR sale_price vs market_value (field 2 vs field 3) |
| 7 | Micromarket Velocity Heatmap (Aishvarya: Intelligence Dashboard map + Our: E3) | Color-coded map showing where transactions are hot or cooling, with drill-through to every deed. Real-time market pulse from government records. | Yes — IGR transactions aggregated by village/taluka + MahaRERA lat/lng |
| 8 | Construction Progress Tracker (Our: E5) | Timeline of construction milestones extracted from RERA quarterly reports — track progress from your phone without a site visit. | Yes — MahaRERA getBuildingWingsActivityDetails API |
| 9 | Comparable Transaction Ping (Our: E1) | Push alert when a new deed registers in your watched micromarket: "2,400 sqft sold at ₹9,200/sqft — 9% above 90-day median." Always-on watchdog. | Yes — IGR daily ingestion + user watchlist |
| 10 | Confidence Thermometer (Our: A2) | Shows "Cap rate 7.8% (confidence: high, 34 transactions)" vs "6.2% (confidence: low, 4 transactions)." The system doesn't pretend all answers are equal. | Yes — transaction count (n) per aggregate, source diversity |
Tier 2 — Buildable with enrichment (GIS, GR, or computed layers)¶
| # | AI Idea | Wow Factor (2 lines) | What's needed beyond RERA+IGR |
|---|---|---|---|
| 11 | Risk Engine / Volatility Model (Aishvarya: Idea 6 + Our: B3) | Multi-axis risk radar (Legal, Policy, Market, Fiscal, Liquidity) with Monte Carlo simulation. Institutional-grade risk view for real estate. | RERA: litigation, complaints, encumbrances. Needs: macro rates, policy event calendar |
| 12 | Location Intelligence (Aishvarya: Ideas 7-10) | Four sub-views: Zone Overview (yield + momentum), Risk Matrix, Zoning/FSI Rules, Traffic Flow — all from one lat/lng. | RERA: lat/lng, FSI. Needs: GIS layers, traffic APIs, zoning data |
| 13 | Live Intelligence Feed (Aishvarya: Idea 5 + Our: C1, E2) | Real-time signal feed: new GRs, infra catalysts, yield shifts, sentiment changes — each linked to source. Bloomberg terminal for real estate. | RERA + IGR for base. Needs: GR scraping, news/sentiment NLP |
| 14 | Cross-Source Contradiction Detector (Our: C2) | Flags when developer website says 85% sold but RERA says 62% and IGR registrations show 41%. Three data sources, three different stories. | RERA: unit counts. IGR: deed counts. Needs: developer website scraping |
| 15 | Stamp Duty Deadline Effect (Our: B4) | Annotates velocity charts: "volume inflated 40-60% before stamp duty deadline — adjust seasonal comparisons." Flags its own data's bias. | IGR velocity data + GR stamp duty concession calendar |
Tier 3 — Requires user data or external feeds¶
| # | AI Idea | Wow Factor (2 lines) | What's needed |
|---|---|---|---|
| 16 | Portfolio AI Advisor + Wealth Path (Aishvarya: Ideas 3, 11) | "Your Path to ₹10Cr" with asset-by-asset Hold/Rebalance/Exit recommendations and 1,000-run Monte Carlo projections. | User portfolio data, macro rates, appreciation models |
| 17 | Alpha Selection / Asset Gallery (Aishvarya: Ideas 4, 12 + Our: D2) | AI-curated marketplace of vetted opportunities with confidence scores, yield targets, and "Why Proppie Recommends" narratives. | Fractional platform listing data + computed scores |
| 18 | Persona-Aware Intelligence (Our: D1, D3, D4, D5) | NRI due diligence packs, tax-bracket-aware yields, family-stage context, ELI5 toggles — the system adapts to who you are. | User self-reported persona data |
Raw Fields → UX Translation (Technical Approach)¶
This is NOT about the pipeline. This is how each raw field becomes a visible AI experience.
Layer 1: Direct Display (raw field → screen element)¶
| Raw field(s) | Transformation | UX element |
|---|---|---|
MahaRERA projectName, projectRegistartionNo |
None — display as-is | Dossier header, search results |
MahaRERA latitude, longitude |
Plot on map | Zone overview pin, heatmap center |
MahaRERA landAreaSqmts, totalPermissiblePlotFsi |
None | Zoning specs panel (FSI/FAR, coverage) |
MahaRERA buildingNameNumber, noOfBuildingWingFloorProposed |
None | Building details card |
MahaRERA complaintDetails[] |
Count + list | Developer score complaints component |
MahaRERA ccDocumentFileName, ccIssuedDate |
Link + date | Approval timeline in dossier |
MahaRERA activities[] |
Map activityId → milestone label | Construction progress timeline |
IGR deed_type (field 1) |
Classify | Transaction type badge |
IGR sale_price (field 2) |
Format as ₹ | Transaction card, comparable table |
IGR registration_date (field 10) |
Format date | Timeline, velocity calculation input |
IGR seller_name, purchaser_name (fields 7, 8) |
Display names | Party details in source trail |
IGR stamp_duty (field 12), registration_fee (field 13) |
Format as ₹ | True cost calculator line items |
IGR party_pan |
Entity resolution key | Cross-deed buyer/seller matching (not displayed) |
Layer 2: Simple Computation (raw fields → derived metric → screen element)¶
| Computation | Input fields | Formula | UX element |
|---|---|---|---|
| Price per sqft | IGR sale_price ÷ IGR area (field 5) |
Direct division (convert sqm→sqft) | Price benchmark, comparable cards |
| ASR Gap % | IGR sale_price (field 2) − IGR market_value (field 3) ÷ field 3 × 100 |
(sale - asr) / asr × 100 |
ASR Gap Scanner badge (e.g., "+110% above ASR") |
| Delay months | MahaRERA projectProposeComplitionDate − today (or actual completion) |
Date diff in months | Delay score bar, developer trust sub-component |
| Transaction velocity | Count of IGR deeds per village per 90-day window | COUNT(deeds) WHERE village=X AND date > now-90d |
Heatmap cell intensity, trending zones sidebar |
| Median price 90d | All IGR sale_price / area per micromarket, last 90 days |
Median of set | Velocity heatmap tooltip, comparable ping threshold |
| Stamp duty % | IGR stamp_duty ÷ IGR sale_price × 100 |
Division | True cost calculator stamp duty line |
| Units registered vs RERA declared | Count IGR deeds for project ÷ MahaRERA totalUnitCount |
Division | Cross-source contradiction flag |
| Buyer origin analysis | IGR party_address → state extraction |
String parsing | Investor demand heatmap (which cities buy here) |
Layer 3: AI/LLM Layer (computed metrics + context → narrative)¶
| AI function | Inputs | Output | UX element |
|---|---|---|---|
| Dossier narrative | All MahaRERA fields + top 10 IGR transactions for project | Executive summary paragraph | Investment Dossier executive summary |
| Risk explanation | Litigation flag, complaint count, delay months, ASR gap | "Zone Risk Index: 3.2/10 — Primary drivers: favorable zoning, low litigation" | Risk Engine panel |
| Comparable context | New transaction + 90d median + percentile | "This sale is 9% above the 90-day median — in the 85th percentile for this micromarket" | Transaction ping notification |
| Cost breakdown narrative | Stamp duty + registration + GST rules + district rates | "Total true cost: ₹99.3L — 17% above listed price. Here's the breakdown..." | True Cost Calculator output |
| Progress assessment | MahaRERA activity milestones + original timeline | "Project is 2 quarters behind schedule. Developer's other projects average 4-month delay at this stage." | Construction progress narrative |
MVP Implementation Sequence¶
Sprint 1 (Weeks 1-3): Data Foundation¶
- Parse IGR Index-II PDFs → structured JSON (Marathi NLP for field 4)
- Ingest MahaRERA APIs → canonical store (20 key endpoints)
- Entity resolution: match IGR deeds to MahaRERA projects via survey number + village
Sprint 2 (Weeks 4-6): Core Screens¶
- Transaction Explorer — searchable table of IGR deeds with drill-through to Index-II fields
- Developer Trust Score — computed from MahaRERA complaint + delay data
- True Cost Calculator — rule-based, district-specific
Sprint 3 (Weeks 7-9): Intelligence Layer¶
- Investment Dossier — per-project report aggregating RERA + IGR data, LLM narrative
- ASR Gap Scanner — per-transaction gap with micromarket percentile context
- Micromarket Velocity Heatmap — village-level aggregation on map
Sprint 4 (Weeks 10-12): Real-Time + Polish¶
- Comparable Transaction Pings — daily IGR ingestion → alert engine
- Construction Progress Tracker — activity milestone timeline from RERA
- Confidence Thermometer — n-count and source diversity on every metric
- UI: dark-mode, gold-accent design system per Aishvarya's DESIGN.md
What's NOT MVP (and why)¶
| Screen | Why not Phase 1 |
|---|---|
| Wealth Path / ₹10Cr projection | Needs user portfolio data + macro models — no point without real holdings |
| Portfolio AI Advisor (Hold/Rebalance/Exit) | Crosses into SEBI IA territory — needs compliance framework first |
| Alpha Selection marketplace | Needs fractional platform listing integration |
| Traffic Flow / Pedestrian heatmap | Needs Google/Mapbox traffic API integration |
| Zoning Rules detail | Needs DCR/DP data digitization — not available in RERA/IGR |
| Live Intelligence Feed (full) | Needs GR scraping + news NLP — add post-MVP |
| Sentiment alerts | Needs social/news corpus + NLP pipeline |
Wow Factor Summary — All 18 Ideas¶
| # | Idea | Wow in 2 lines |
|---|---|---|
| 1 | Transaction Explorer + Source Trail | Every price claim links to the actual government deed. Users see arithmetic over registered data, not broker guesswork. |
| 2 | Investment Dossier | One-click deep-dive report with RERA details, cost breakdown, risk flags, and comparable transactions — all source-cited. Replaces weeks of manual research. |
| 3 | Developer Trust Score | A transparent 0-100 score where every bar links to the RERA filing. Users can disagree with weights but can't dispute the inputs. |
| 4 | True Cost Calculator | Reveals the 17% hidden cost above "listed price" — stamp duty, GST, maintenance, parking, society transfer. Jaw-drop moment for first-time buyers. |
| 5 | Carpet Area Reality Check | Catches brochure inflation instantly using RERA's own filings. Exposes the industry's oldest trick with government data. |
| 6 | ASR Gap Scanner | Flags when a sale price is abnormally above or below the government's own valuation. Instant overpricing/distress signal nobody else provides. |
| 7 | Micromarket Velocity Heatmap | Color-coded map showing exactly where transactions are hot or cooling, with drill-through to every deed. Real-time market pulse from registered data. |
| 8 | Construction Progress Tracker | Track building progress from your phone using only RERA quarterly reports — no site visit, no broker, no developer spin. |
| 9 | Comparable Transaction Ping | Always-on watchdog: alerts when a new deed registers in your market that's above or below the running median. |
| 10 | Confidence Thermometer | The system admits when it doesn't have enough data. "34 transactions = high confidence" vs "4 transactions = treat with caution." |
| 11 | Risk Engine / Volatility Model | Institutional-grade multi-axis risk radar with Monte Carlo simulation — brings hedge fund analytics to Indian real estate. |
| 12 | Location Intelligence (4 views) | Four sub-views from one lat/lng: yield forecast, risk matrix, zoning/FSI, traffic — a Bloomberg terminal for real estate locations. |
| 13 | Live Intelligence Feed | Real-time signal processing: GR impact, infra catalysts, yield shifts. The platform watches government portals so you don't have to. |
| 14 | Cross-Source Contradiction Detector | Catches developer lies: website says 85% sold, RERA says 62%, IGR registrations show 41%. Three sources, three stories, one truth. |
| 15 | Stamp Duty Deadline Effect | Flags its own data's seasonal bias — "volume inflated before stamp duty deadline." An analytics tool honest about its own blind spots. |
| 16 | Wealth Path + Portfolio Advisor | "Your Path to ₹10Cr" with Monte Carlo projections and Hold/Rebalance/Exit signals per asset. Financial planning meets real estate. |
| 17 | Alpha Selection / Asset Gallery | AI-curated marketplace where every listing has a confidence score, yield projection, and "Why Proppie Recommends" — vetted, not listed. |
| 18 | Persona-Aware Intelligence | NRI packs with FEMA rules, tax-bracket-aware yields, family-stage context. The system adapts to who you are without you asking. |
Cross-references:
- Raw field inventory: ../20-data/source-raw-fields.md
- Field mapping: ../20-data/source-field-mapping.md
- Aishvarya's original ideas: aishvarya-ux-ideas.md
- Detailed wow moments: wow-moments-catalog.md
- Data attributes: ../20-data/data-attributes.md