Property Search and Advisory Copilot
Outcome
Help a buyer, tenant or broker narrow a property search using verified listing attributes, neighborhood and geospatial context and financial feasibility, while keeping every statement inside fair-housing bounds and grounded in disclosed listing data rather than inference about occupants.
| Portfolio attribute | Value |
|---|---|
| Primary persona | Buyer, tenant, or broker |
| Primary aggregate | PropertyListing |
| Cosmos DB containers | 2 |
| Grounding data | listings, amenities, geospatial and market data |
| Development estimate | USD 60-220 monthly dev estimate |
| Runtime options | Python FastAPI and C# ASP.NET Core |
Productivity hypothesis
These measures reflect the workflow, evidence, decision boundary, and industry risk for this use case.
| Impact type | Measure | Conservative internal assumption | Customer measurement |
|---|---|---|---|
| Productivity | Shortlist and comparison time | 20-35% reduction | (baseline median property-search minutes - assisted median minutes) / baseline median minutes * 100 |
| Quality or risk | Accepted recommendations | 5-10 percentage-point improvement | assisted percentage of recommendations accepted - baseline percentage |
Important: Conservative internal pilot assumptions only; not public benchmarks, commitments, or expected customer results. Validate against a comparable baseline or matched control cohort with mandatory human review.
Rationale: Buyers, tenants, and brokers can compare listings and market evidence faster, subject to current data coverage.
Pilot guardrail: Disclose listing, availability, price, amenity, geospatial, and market-data freshness limitations.
Customer validation plan
- Baseline the current workflow with representative users and tasks.
- Run the accelerator on the same task mix with human review retained.
- Measure shortlist and comparison time and accepted recommendations with the documented formulas.
- Replace internal assumptions with customer-observed results before a business case.
Why Azure Cosmos DB
This AI solution needs an operational data plane, not only a vector index:
- Hybrid vector and full-text retrieval combines semantic recall with exact identifiers, codes, names, and domain language.
- Hierarchical partition keys and metadata predicates keep tenant and domain boundaries in the normal query path.
- Operational JSON, source chunks, workflow and agent state, feedback, and audit records remain together for low-latency reads and governed writes.
- Change feed reacts to source and operational updates without polling.
- Autoscale throughput, configurable consistency, and multi-region distribution support variable demand and enterprise availability.
Domain workflow

| Stage | Actor | Input → output | Approval |
|---|---|---|---|
| 1. Declared requirement capture | Buyer, tenant or broker | clientConversation → BuyerRequirementProfile | Policy controlled |
| 2. Criteria-based candidate retrieval | Property search and advisory copilot | BuyerRequirementProfile → candidateListingSet | Policy controlled |
| 3. Comparable and price positioning | Valuation analyzer | MarketComparable → pricePositionAnalysis | Policy controlled |
| 4. Fair-housing compliance screening | Compliance guardrail service | AdvisoryBrief → fairHousingCheckResult | Policy controlled |
| 5. Broker review before client delivery | Licensed broker | AdvisoryBrief → deliveredAdvisory | Human required |
Technical architecture

Cosmos DB data model
| Container | Hierarchical partition key | Documents | Search paths |
|---|---|---|---|
propertyListings | /tenantId, /listingId | Listing, ListingDetail, MarketComparable, BuyerRequirementProfile, AdvisoryBrief | /narrativeVectorEmbedding, /advisoryVectorEmbedding, /narrativeVector, /advisoryVector |
neighborhoodFactLibrary | /tenantId, /geographyCode | NeighborhoodFactChunk | /chunkVectorEmbedding, /chunkVector |
Query contract
| Query | Container | Scope | Purpose |
|---|---|---|---|
| CandidateListingsByDeclaredCriteria | propertyListings | cross-partition-bounded | Return active listings matching the client's declared must-haves and budget in the target market areas. Intentional cross-partition listing search: bounded by the tenantId prefix, active status, the buyer's target market areas and declared price and bedroom limits. |
| ComparablesForPricePosition | propertyListings | single-partition | Pull recent comparable closings near the listing to explain where its asking price sits. |
| PermissibleNeighborhoodFacts | neighborhoodFactLibrary | single-partition | Retrieve only neighborhood facts classified as fair-housing permissible for the question asked. |
| AdvisoriesAwaitingBrokerReview | propertyListings | cross-partition-bounded | Queue drafted advisories that a licensed broker must review before delivery. Intentional cross-partition broker review queue: bounded by the tenantId prefix, the pending review state and a cleared fair-housing check. |
Python and C# services route each document type to its declared container and hierarchical partition key. ETag-based writes protect mutable aggregates, while append-only records and bounded cross-partition queries preserve scale.
Azure deployment

Clone the source repository, authenticate with Azure, and run:
Set-Location accelerators/real-estate-construction/property-search-and-advisory-copilot
az login
./Deploy.ps1 -Runtime python -Location eastus2
Change python to csharp to deploy the .NET implementation.
Explore the implementation
- Launch signed-in demo
- Implementation article
- Domain data model
- Workflow and API contract
- Python application
- C# application
- Domain container Bicep
- Platform Bicep
- Cost and configuration
- Source repository