# From dozens of sources to one curated event calendar

Source: ArrowEye (https://arroweye.be) · Language: English · Canonical page: https://arroweye.be/en/cases/eventagenda-networkingplatform

A scrape-and-enrich chain that collects, filters and deduplicates B2B networking events from dozens of sources; the team approves every list before it goes online.

## Facts

- **Category:** Transform (discovery, agents and automations)
- **Sector:** Networking platform · Startup
- **Technology:** n8n

## The question

An up-to-date calendar of B2B networking events in Belgium and the Netherlands on their own platform, without the team manually trawling dozens of event platforms, employer federations, exhibition halls and networking clubs every week. The solution had to tell business events from consumer and online events, keep only events with a confirmed date and a physical venue, recognise duplicates across sources, and put every list to the team first before anything went online.

## Our approach

A chain that reads the main event sources periodically, including the sources that do not give up their data easily. Every event is tested against a criteria document fixed in advance: geography, physical venue, time window and business signals. AI enriches what remains with event type, organiser, sector tags and a usable description, and recognises the same event across different sources. Every new list is ready as a proposal and only goes to the website after the team approves it. Past events are cleaned up automatically, so the calendar stays current.

## Impact delivered

From scope to working MVP in a few weeks, then handed over to the client's own environment with a handover document, so the team builds on independently. The weekly searching and retyping disappears; events arrive in one fixed structure, from date and venue to sector tags and registration link. Quality control stays with the team, not with the scraper. The chain also showed which sources really deliver volume, and can be extended to more countries and sources without a rebuild.

## The route

1. scrape of event sources
2. filtering against a criteria document
3. AI enrichment and deduplication
4. proposal for approval
5. publication on the website