SponserdIQ

Enrichment Pipeline

SponserdIQ uses an AI-powered enrichment pipeline to augment scraped sponsorship data with structured company information and deal intelligence.

Two-Stage Enrichment

Brand Enrichment

When a new brand is identified, SponserdIQ enriches it with company-level data:

Field Description
company_size Company scale (e.g., startup, mid-market, enterprise, public)
revenue_range Estimated revenue band (e.g., $10M-$100M, $100M-$1B, $1B+)
founded_year Year the company was founded
industry Primary industry classification
description Brief company description

Enrichment only fills in fields that are currently empty. If a staff member has manually set a value (e.g., corrected an industry classification), the AI will not overwrite it.

Sponsorship Enrichment

Each sponsorship relationship is enriched with deal-level intelligence:

Field Description
activation The sponsorship activation description (e.g., Official Healthcare Partner, Naming Rights Partner)
sponsorship_type The sponsorship type (naming_rights, presenting, title, official, signage, digital, community)

Sponsorship enrichment uses context text from the original scraped source (the raw sponsor page content) to make informed classifications.

How Enrichment Triggers

Enrichment is triggered in three ways:

1. After Brand Classification

When a raw sponsor record is successfully classified into a brand, the system automatically queues both brand enrichment and sponsorship enrichment. This means new brands and sponsorships are enriched shortly after they are first discovered.

2. Scheduled Enrichment

An automated process runs on a recurring schedule after scrape cycles complete. It finds brands and sponsorships that have not yet been enriched and queues enrichment jobs for them in controlled batches.

3. Manual Admin Action

Staff can trigger enrichment for individual brands or sponsorships from the admin panel. This supports a force option that bypasses the cooldown period.

Cooldown and Deduplication

  • Enrichment has a cooldown period -- recently enriched records are skipped unless the force option is used.
  • Deduplication logic prevents the same record from being queued for enrichment multiple times within the cooldown window.
  • Enrichment jobs include automatic retries to handle transient failures.

Enrichment Indicators

In the SponserdIQ dashboard, enriched records are marked with a sparkle indicator. This tells users that the data includes AI-generated fields beyond what was directly scraped.