Loading
Internal Product · In Development
A production-grade lead generation, opportunity qualification, and BD platform - automated ingestion, two-layer AI classification, 8-dimension scoring, and AI-generated briefs.
The Opportunity Engine automates the discovery, classification, scoring and briefing of commercial opportunities from UK procurement portals, startup signals and tech-news sources - surfacing only the leads that map to Indigo Pascal's service lines. It's the BD nervous system of the studio.
Scheduled crawlers pull from Find a Tender, Contracts Finder, BidStats, UK tech RSS feeds and curated startup signals - every relevant opportunity surfaces automatically.
Deterministic keyword engine for speed and explainability, with a GPT-4o-mini fallback for ambiguous edges - across 12 Indigo Pascal service categories.
Weighted model: service fit, geography, value, urgency, complexity, conversion signals, repeat potential, and project type - every lead gets a defensible score.
Rule-based + LLM-powered briefs covering scope, budget band, risks, qualifying questions and recommended next steps - ready for the BD team in seconds.
Source ID + weighted text similarity at a 0.75 threshold - no double-counting across portals.
Detected → triaged → qualified → briefed → engaged → bidding → won/lost/archived, with full audit trail.
KPIs, pipeline charts, source performance, geographic split and service-line breakdowns.
5 roles - admin, director, BD manager, BD user, viewer - backed by JWT auth.
Pluggable adapter framework so a new procurement portal or signal source is days, not weeks.
Every LLM path has a rule-based shadow - the engine works even when the API key is offline.
Pluggable adapter framework - every source normalises into a common Opportunity model so dedup, classification and scoring stay consistent across feeds.
Every weight is configurable. Every score is explainable. The same model that ranks opportunities is the one we tune against win/loss feedback - closing the loop on what actually converts.
I've been working on Aceternity for the past 2 years. Here's a timeline of my journey.
Adapters, deterministic classifier, dedup, seed data, JWT auth.
8-dimension scoring, rule-based briefs, LLM fallback wiring.
Full 8-stage pipeline, dashboard KPIs, geographic + source charts.
Docker compose, RBAC, audit trail, alembic migrations, pytest coverage.
Templated outreach, signal-driven nudges, CRM sync, win/loss feedback loop.
Request a walkthrough
Live demo of ingestion, classification, scoring, and AI briefs - tuned to your sector before the call. We will send a recording and a tailored capability statement after.