From monolith to microservices: preparing
an AI-native investment platform for launch 

Investment management for family offices is a domain defined by complexity. These organizations manage exposure across private markets, alternative assets, and multi-entity structures, all of which require operational precision and real-time visibility that generic financial tools cannot provide.

The client set out to build the first AI-native platform purpose-designed for this space. But when Honeycomb Software joined the project, the product faced significant technical challenges that needed to be resolved before a public launch was possible.

The platform had been built as a monolith — a single, tightly coupled codebase that made it difficult to scale individual components, introduce new features safely, or maintain reliable performance under growing load. As the product’s ambitions grew, the architecture became a bottleneck.

Key challenges included:

• Monolithic architecture: the existing codebase could not scale to support the platform’s roadmap without a fundamental restructuring.
• Frontend performance and maintainability: the frontend needed to be rebuilt to support a modern, responsive user experience suitable for a demanding financial audience.
• Feature delivery under time pressure: new capabilities had to be developed and shipped alongside the architectural overhaul, without delaying the planned launch
• Complex integrations: the platform required deep integrations with financial data sources, document processing, email services, payment infrastructure, and Microsoft ecosystem tools.
• AI-powered document workflows: the system needed to classify and process investment documents intelligently, using large language models as part of the core product experience.

Honeycomb Software took on the full technical transformation of the platform, rearchitecting the backend, rebuilding the frontend, and delivering new features in parallel, all within the timeframe required for a December 2025 launch.

From monolith to microservices

The existing monolithic backend was decomposed into a microservices architecture, with each service independently deployable and scalable. This restructuring gave the platform the flexibility to evolve quickly, absorb new integrations, and handle growing usage without systemic risk.

The backend was built on Nest.js with TypeScript, using Drizzle ORM and TypeORM for data management, and deployed across Azure Web Apps with Azure Service Bus handling asynchronous messaging between services.

Frontend rebuild

The frontend was rebuilt using Next.js and ReactJS with TypeScript, delivering a fast, maintainable interface suited to the complexity of investment management workflows. TanStack Query was used for efficient data fetching and caching, Recharts for financial data visualisation, and Tailwind CSS for a clean, consistent UI. The result was a frontend that could keep pace with the platform’s expanding feature set.

New feature development

Alongside the architectural work, Honeycomb Software developed new platform capabilities, including:

• Document ingestion: users can upload PDFs directly to the platform or forward emails with attachments to a personal Unlimited.ai inbox. The system automatically recognises document types — from capital calls and distribution notices to LPAs, K-1s, and financial statements — with no manual tagging required.
• AI-powered document processing: once a document is received, the platform’s AI engine parses and classifies it automatically, extracting key data points such as fund names, capital committed, dates, wire instructions, carry terms, and entity-specific details — with no manual intervention required.
• Natural language portfolio Q&A: users can query their investment data in plain language and receive precise, source-cited answers drawn directly from their documents and portfolio history — turning the document archive into a live, searchable knowledge base.
• Portfolio and performance tracking: the platform automatically updates the investment ledger as new documents arrive, providing real-time tracking of IRR, MOIC, DPI, net cash position, and capital flow timelines — viewable per investment, fund vintage, asset class, or entity.
• Smart document management and search: all documents are automatically tagged by type, indexed for search, and linked to relevant investments. Users can search by keyword, fund name, timeframe, or document type, and run contextual queries across the full document library.
• Multi-entity reporting: the platform supports multi-entity account structures — LLCs, trusts, family partnerships, and holding companies — with per-entity performance tracking, roll-up reporting across entities, and K-1 consolidation for tax season readiness.

The platform launched successfully in December 2025. The architectural transformation and feature delivery completed by Honeycomb Software positioned the product for a stable, scalable market entry, moving from a pre-launch monolith to a production-ready, microservices-based platform within the planned timeline.

The client acknowledged Honeycomb Software’s contribution to the launch, describing the outcome as a success. The delivered platform now provides the technical foundation for the client’s broader vision: redefining what is possible for family office investment management through AI-native tooling and autonomous portfolio intelligence.