Bilingual Property Portal for an Auckland Agency
A bilingual property portal with AI-drafted listings, unified buyer CRM, and an on-prem deployment option for agencies that want data to stay in NZ.
- Client
- Sample scenario: Auckland real estate agency
- Industry
- Real Estate Agency
- Services
- AI Web DesignEnterprise SoftwareAI Integration
- Listings drafted in both EN and 中文 without a separate copywriter
- All inquiries (site, email, messaging) consolidated into one CRM view
- Option to deploy on the agency's own server — data never leaves NZ
The scenario
Auckland's Chinese buyer market is significant, but most real estate platforms serve English speakers first and translate the Chinese version poorly (or not at all). A typical 20-agent agency loses inquiries from Chinese-speaking buyers because:
- Listings are English-only, or the Chinese version is a thin machine translation.
- Each listing takes 30-60 minutes to write manually.
- Inquiries come through multiple channels (website form, email, messaging apps, phone) and live in nobody's CRM.
This case study describes what we can build remotely for an agency with this shape.
What we build
A bilingual property portal that:
- Drafts listings in both languages. Agents upload photos and structured details (address, bedrooms, features); Claude drafts the English and Chinese listing copy in the agency's voice, which the agent edits and publishes. Copy quality is comparable to a mid-level bilingual copywriter — because that is, effectively, what the model was trained to emulate.
- Consolidates inquiries into one CRM. Every inquiry — website form, email, WhatsApp — lands in a single Kanban-style view with buyer history, saved listings, and conversation log. (Note: WeChat Mini-Programs require a China-registered entity and ICP filing, so we don't claim to build those. We do build mobile-optimized landing pages that render correctly when shared in WeChat's in-app browser, with proper share-card metadata.)
- Matches buyers to listings automatically. A lightweight classifier compares new listings against saved buyer preferences (area, budget, bedrooms, features) and sends the agent a daily suggested-followups list.
- Publishes a properly bilingual public site. Fast, SEO-optimized in both languages with correct
hreflang, Google Maps embeds, and structured data so listings appear correctly in Google's real estate rich results.
Deployment options
- Cloud (default): we host on our infrastructure. Fastest to launch, no infra overhead for the agency.
- On-premises / own server: we ship the system as a Docker Compose stack the agency installs on its own office server or private cloud. All buyer data stays inside the agency. Useful where data-sovereignty or a privacy-conscious principal is a factor.
Both options are fully source-owned by the agency — you get the code repo.
Pricing
Scope-quoted. After a free 30-minute consultation we produce a written fixed-price quote — we don't do hourly billing with a vague estimate. Delivery is typically 3-6 weeks depending on the integrations you need.