New State of AI 2026: Mid-Year Reality Check is live. Read the report
Skip to content
Case Study Commercial Real Estate

From AI Strategy to a Production Multi-Agent AI Platform

Secure, Conversational Access to Enterprise Knowledge and Deal Workflows for a National Commercial Real Estate Firm

Client National commercial real estate firm
Market United States
Cloud Google Cloud Platform
A national commercial real estate firm using a multi-agent AI platform for enterprise knowledge and deal workflows
Hatchworks Team AI/ML Engineers, Solution Architect, Data Engineers, Product Owner, Project Manager
Overview

An intelligence layer on the cloud the firm already runs

HatchWorks AI first partnered with a national commercial real estate firm on an AI strategy and roadmap engagement, then moved into building a production-oriented multi-agent AI platform focused on high-value real estate workflows. The platform launched with three capabilities: an enterprise knowledge assistant, broker document generation, and title and lease abstraction.

Delivered on the client’s existing Google Cloud stack, the platform provided an intelligence layer giving teams secure, natural language access to enterprise knowledge and deal data without duplicating data or weakening security controls. It established a reusable foundation for future AI use cases across the business.

Knowledge assistant

Natural language Q&A over enterprise knowledge, grounded in source systems.

Document generation

Broker-facing BOV and Offering Memorandum creation.

Title and lease abstraction

Structured data extracted from title and lease documents.

The Challenge

Unlock enterprise knowledge without moving the data

The firm needed to unlock enterprise knowledge and deal-related data spread across multiple systems, without copying that data into a new store or weakening security. Teams relied on manual, fragmented workflows to answer internal questions, produce broker-facing documents such as Broker Opinions of Value (BOV) and Offering Memoranda (OM), and pull structured information from title and lease documents.

The solution also had to respect strict architectural constraints:

  1. 01

    Reuse the firm’s existing Google Cloud stack wherever possible.

  2. 02

    Keep sensitive data inside the approved environment.

  3. 03

    Support role-based access and operational compliance.

  4. 04

    Deliver useful functionality quickly, rather than waiting for a large, risky big-bang rollout.

The Process

Strategy first, then three capabilities in phased releases

HatchWorks AI began with stakeholder interviews, workshops, a readiness assessment, and opportunity prioritization to anchor the work to measurable business value. That strategy phase produced a phased roadmap and the three initial use cases to build first.

From there, HatchWorks and the client worked in an integrated delivery model to design and implement the platform.

  1. 01

    Secure integrations

    Connected to the firm’s content management, CRM, and productivity systems through secure SaaS integrations, using authentication and token-vault patterns.

  2. 02

    Enterprise knowledge assistant

    Built a natural language assistant for enterprise knowledge access, grounded in source systems through retrieval-augmented generation.

  3. 03

    Broker document generation

    Delivered document generation workflows for BOV and OM creation.

  4. 04

    Title and lease abstraction

    Built OCR, classification, and extraction pipelines to abstract structured data from title and lease documents.

Delivery followed iterative, sprint-based releases with user feedback loops, plus production hardening across authentication, guardrails, observability, and data access.

How a question reaches an answer

Source systems stay where they are. The intelligence layer reads them in place.

Step 01 Source systems
Content management CRM Productivity systems Title and lease documents Web search
Step 02 Secure integration layer
Authentication and token vault Secure SaaS integrations read source systems in place. No data duplication.
Step 03 Intelligence layer on GCP
Retrieval-augmented generation Answers grounded in source systems, with Gemini and other LLM services.
Guardrails and observability Role-based access control, tracing, and production hardening.
Step 04 Three agents
Knowledge Q&A BOV and OM generation Title and lease abstraction
Step 05 Teams
Natural language access Research, internal knowledge, broker deliverables, and document workflows, under role-based access.
The Outcome

A working platform, and a foundation to build on

The engagement moved the firm from AI strategy into a working multi-agent Intelligence Platform and shipped the initial release, creating a reusable foundation for future AI-powered workflows. The work supported rollout across the business, enabled secure natural language access to internal systems, and established delivery patterns for document generation and structured extraction.

The initiative gave the firm real momentum: teams saw their AI progress accelerate, and HatchWorks AI was a hands-on partner throughout the rollout. The platform positioned the firm to keep expanding into additional agents and deeper integrations, rather than treating AI as a one-off pilot.

Enterprise knowledge and deal data spread across multiple systems.

Manual, fragmented workflows to answer internal questions.

BOV and OM documents produced by hand.

Structured information pulled from title and lease documents by hand.

Key Stats

The program in numbers

3

initial use cases delivered: enterprise knowledge assistant, BOV composition, and title and lease abstraction

24

weeks planned across three phased releases

70+

users in production, with rollout continuing in waves of roughly 20 and 30 users

No data duplication

secure, natural language access to enterprise systems

Phased delivery Planned windows, three releases.
Core platform 10 weeks
Document generation 8 weeks
Abstraction 6 weeks

Planned as a 24-week program across three phased releases, delivered in iterative sprints with user feedback loops.

Technologies Used

The stack

Cloud and models
Google Cloud Platform (GCP) Gemini and other LLM services Retrieval-augmented generation (RAG) Web search integration
Integrations and security
Secure SaaS integrations across content management, CRM, and productivity systems Authentication and token-vault patterns
Document intelligence
OCR and document classification and extraction pipelines PDF and document generation workflows
Operations
Observability and tracing for production support
Context

About the work

About the client

The client is a national commercial real estate firm operating in the United States, pursuing a broader AI transformation agenda across research, internal knowledge access, broker deliverables, and document workflows.

About HatchWorks AI

HatchWorks AI turns AI into ROI by automating key business processes, transforming data, deploying intelligent agents, and shipping AI-powered products that deliver measurable results.

Ready to build the intelligence layer your business runs on?

See how HatchWorks AI puts agents to work inside the systems your teams already use.