Write The Future

Problem: You hire, you train, they quit, you write off the wasted resources as expense and start over. Solution: An AI platform that can record talent capital as a new asset class on balance sheets.

Comment

Inspiration

We were inspired by a simple but powerful question: "What if companies could hire AI employees that work, learn, and collaborate like humans?" Even better, "What if companies can monetize their well trained work forces and rent them out in a marketplace." That vision became Write The Future, an AI platform where businesses build, train, and collaborate with intelligent digital employees.

What it does

Write The Future allows businesses to:

  • Register and set up their company space with full branding (logo, colors, policies).
  • Generate AI departments and teams (e.g., Tech, Operations, Marketing), each led by AI executives like CTOs or COOs.
  • Train AI employees using real company knowledge — documents, PDFs, videos, or internal APIs.
  • Create and automate workflows using BPMN-like process maps and Genkit flows.
  • Collaborate in real time through an embedded Kanban board where both humans and AI employees manage, discuss, and deliver tasks.
  • Integrate MCP servers and APIs from a marketplace of 100+ third-party services or custom company APIs.

In short: it’s a living, evolving AI organization inside your organization that makes AI transformation feels like a breeze.

How we built it

We built Write The Future using a stack that combines modern web technologies with scalable AI orchestration:

  • Frontend: Angular (standalone architecture) + Angular Material + SSR for SEO and performance.
  • Backend: GCP, Firebase (Firestore, Functions, Auth, Storage, Eventarc) + Genkit for AI flow orchestration.
  • AI Layer: Custom Genkit agents communicating via Agent-to-Agent (A2A) and Model Context Protocol (MCP) for tool calling and external API access.
  • Integration Layer: API marketplace with dynamic connectors MCP servers as well.
  • Realtime Collaboration: Firestore streams with signal-based state management for low-latency updates.

We used a modular workspace where a shared folder houses reusable services, models, and themes. Each agent type (AI employee) is deployed as a composable flow, capable of reasoning, planning, and conversing autonomously.

Challenges we ran into

  • Multi-agent synchronization: Ensuring multiple AI employees coordinate without redundant or conflicting actions required designing a lightweight orchestration layer inspired by distributed consensus models.
  • Knowledge ingestion: Parsing and embedding multi-format documents (PDFs, DOCs, videos, audio) while preserving context was complex.
  • Dynamic SEO for SSR: Making AI-generated pages SEO-friendly while rendering dynamically via server-side Angular was technically challenging.
  • Realistic human–AI collaboration: Designing UI/UX patterns that make interacting with AI employees feel natural — not robotic — required iterative refinement.
  • Resource management: Balancing LLM inference costs and Firebase scalability for hackathon constraints.

Accomplishments that we're proud of

  • CoinFest Asia 2025 (August 2025): Earned recognition as the most innovative AI solutions (1st Prize winner in AI startup battle).
  • NVIDIA Inception (September 2025): Accepted into the program and granted access to $1M+ in grants, plus the marketing assets and NVIDIA VC Network (150 VCs).
  • Google Ecosystem Partner (October 2025): Accepted into Google Ecosystem and granted $25,000 in cloud credits.

What we learned

We learned that building AI organizations is less about automation and more about coordination — aligning AI reasoning with real business intent.

We also gained deeper insights into:

  • AI memory management and contextual grounding.
  • The importance of structured data (JSON-LD, metadata) for discoverability.
  • Creating feedback loops where AI agents improve over time based on task outcomes.

And most importantly, we learned that AI collaboration can feel human when done right.

What's next for Write The Future

Thanks to the feedback from our early adopters, we’re expanding Write The Future into a B2B AI platform where any company can:

  • Onboard instantly and create their AI organization.
  • Train agents on internal policies and tools.
  • Deploy AI workflows via Genkit or custom APIs.
  • Visualize real-time progress through dashboards and process maps.

In our beta, we onboarded 27 businesses so far of which 8 are ready to subscribe to paid plans as soon as the production is ready (~10 December, 2025)

Built With

Share this project:

Updates