How this dashboard stays current

Scheduled scripts scan each entity's official procurement portal using the playbook's search vocabulary, then normalise and de-duplicate results into the pipeline below. This mockup shows the end state with sample data.

1. ScanA scheduled job queries each of the 9 priority portals + National Treasury eTenders using capability search terms.
2. ExtractTitle, reference, type (RFI/EOI/RFP/RFQ), published & closing dates, and source URL are parsed from each listing.
3. ClassifyEach tender is tagged to A&M capabilities and scored for fit against the core vocabulary.
4. De-duplicateCentral (eTenders) and entity-portal records are merged so the same tender appears once.
5. PublishResults are written to the dashboard; new items since the last scan are flagged.
6. AlertHigh-fit or closing-soon opportunities can trigger an email/Telegram notification (planned).

Scheduling target: hourly for eTenders, daily per entity portal. Backend scrapers are not part of this mockup.