Case study · Real estate and land zoning
Lander.
An AI-powered chatbot for the real estate industry that simplifies land and zoning regulation queries.
- What we built
- AI chatbot web app
- Stack
- Next.js 15.3.1, React 19.0.0, TypeScript 5, Material-UI (MUI) 5.15.11
- Services
- AI chatbotWeb appData processing
About the project
What is Lander?
We built an AI-powered chatbot solution for the real estate industry to simplify land and zoning regulation queries. The platform enables users to enter a district and plot number and instantly receive details about applicable development plans, building permissions, and zoning regulations. Designed for a Polish client, the system integrates local zoning data with legal frameworks, turning static spreadsheets into an intelligent, interactive assistant.
The problem
What was slowing them down.
- 01
Data stored in Excel sheets
The client had land and zoning details stored in Excel sheets, including zoning plans, district data, and area designations.
- 02
Manual effort and cross-checking
Accessing and interpreting this data required manual effort, deep familiarity with Polish building law, and time-consuming cross-checking against government regulations.
- 03
Hard to know what could be built
This made it difficult for property owners, investors, and legal professionals to quickly determine what could be built on a given plot of land.
The solution
What we built.
For property owners, investors and legal professionals
We developed a chatbot powered by GPT-4.1, combined with a structured data backend, to provide real-time answers about plots and zoning.
- Input a district and plot number to view relevant zoning regulations.
- Understand which development plan applies and what construction is permitted.
- Ask follow-up questions about regulations, laws, or building rules in natural language.
- The platform merges a Next.js frontend, AI embeddings, a vector database (Upstash Vector), and document parsing tools.
Results
What changed.
Users and client
Faster decision-making
Reduced time from hours of manual checking to instant chatbot responses.
Improved accessibility
Non-technical users can now understand complex zoning laws with ease.
Scalability
Capable of handling large datasets of plots and zoning plans.
Trust & accuracy
Provides context-driven, regulation-backed answers that align with Polish law.
Competitive advantage
Positioned the client as a digital innovator in the real estate industry.
Key challenges
The hard parts, and how we solved them.
Unstructured Data: original plot and zoning data existed in raw Excel format.
Processed and normalized Excel data into structured JSON for reliable AI querying.
Complexity of Legal Regulations: Polish building laws and zoning rules are highly detailed.
Used embeddings and vector search to ensure the chatbot retrieves precise, context-specific regulations.
Accuracy & Trust: misinterpretation of building laws could cause serious risks.
Integrated AI with regulation-backed document search to provide verifiable answers instead of relying solely on generative output.
User Experience: needed to balance technical accuracy with an intuitive interface.
Designed a clean, modern UI with real-time streaming responses, markdown rendering, and smooth interaction flow.
Tech stack
What it runs on.
- Frontend
- Next.js 15.3.1React 19.0.0TypeScript 5
- UI & Styling
- Material-UI (MUI) 5.15.11Tailwind CSS 4React Markdown 10.1.0React Spinners 0.16.1
- Backend & API
- Next.js API RoutesStreaming responses
- AI & Document Processing
- OpenAI GPT-4.1OpenAI EmbeddingsUpstash Vector
- Data Management
- JSON databaseExcel processing (xlsx 0.18.5)PDF processingRedis 4.7.0
Process
How we got there.
01Data Ingestion
Extracted zoning and plot details from Excel sheets into structured JSON.
02AI Integration
Embedded zoning and legal documents using OpenAI embeddings and stored them in Upstash Vector.
03Frontend Development
Built a modern UI with Next.js, React 19, MUI, and Tailwind CSS for smooth interactions.
04Backend & APIs
Created serverless APIs via Next.js for chatbot queries and data retrieval.
05AI Chatbot
Integrated GPT-4.1 with real-time streaming responses for conversational interaction.
06Testing & Optimization
Ensured accuracy, scalability, and compliance with regulations.
More work.
All projects
Bludeals
AI-powered conversational sales assistant that helps users discover mobile deals, qualify credit, and cross-sell insurance via web and WhatsApp.
HOOPWRLD
AI-powered SaaS platform that evaluates school basketball players, generates detailed performance reports, and tracks progress for players and coaches.
Building something like Lander?