View all work
Shipped

Dealflow Intelligence Platform

AI pipeline that ingests pitch decks from email, parses attachments and produces structured startup intelligence — built to cut manual deal processing.

Dealflow Intelligence Platform
Client
Delivered for a venture capital workflow
Industry
Financial
Category
AI
Year
2024
The challenge

What wasn’t working

VC teams were drowning in inbound decks across inboxes and shared drives — duplicate work, slow triage and no single intelligence trail. Analysts spent hours copying founder data into trackers while partners waited on judgment-ready summaries. We were engaged to turn deal intake into a production pipeline with clear failure modes, not another fragile chatbot prototype.

What we built

Our answer

We built NestJS microservices with GPT-4 document analysis, Gmail/Drive OAuth flows, PDF/DOCX/PPTX parsing, web intelligence (SERP/Apify) and structured JSON outputs with webhooks for downstream tools. Every stage logs what ran and what failed so the firm can audit the trail. Dockerised services keep intake, enrichment and delivery modular as deal volume grows.

Results

What changed

The platform automates the bulk of manual deal-processing workflows: decks reviewed ~10× faster and ~90% of repetitive intake steps automated. Partners spend time on judgment instead of inbox archaeology, with logs teams can inspect when something needs a human eye. Intake stopped being the bottleneck between a cold email and a partner conversation.

Capabilities in play

How it was built

Capabilities in play: this engagement showcases AI & Intelligent Automation and AI & Machine Learning — agents, document intelligence and RAG-style enrichment over pitch materials. We paired APIs & Integrations (Gmail, Drive, webhooks) with DevOps & Infrastructure (Dockerised services) and Security through auditable logs so every automated decision leaves a trail.

Similar challenge?

Tell us what needs solving.

View all work