SecondAct Knowledge Sources V2

4 decisions to advance Digital Jay's knowledge base | March 9, 2026

These decisions shape how we process 7+ new content sources that have arrived via email. Your responses unblock the engineering work — the pipeline infrastructure is built and ready to run.

0 of 4 answered

1. AI Tuesday Columns — Knowledge or Content?

AI Tuesday columns (like the Pomelli piece) and Second Act Spotlight editorials arrive via email. Should Digital Jay be able to reference these in conversations, or are they strictly for newsletter distribution?

A) Knowledge Only

Digital Jay can discuss column topics but we don't track them for distribution.

B) Content Only

Track for distribution cadence but Digital Jay doesn't reference them.

C) Both

Ingest for Digital Jay's knowledge AND track for content distribution. Jay can say "In a recent column, I discussed how Pomelli uses AI for..."

Additional notes (optional)

2. Pull Quotes (100+) — Individual or Grouped?

The pull quotes DOCX contains 100+ standalone quotes from the book and interviews. We need to decide how to chunk them for Digital Jay's retrieval.

A) Individual Chunks

One quote per retrievable unit. Best for targeted retrieval: "give me a quote about reinvention." Theme tags added as metadata.

B) Grouped by Theme

Cluster related quotes together. Fewer chunks, but less granular search. Better if you want Digital Jay to present themed collections.

Additional notes (optional)

3. Manuscript Re-ingestion

The original 341 manuscript chunks use an old schema without content_type or source attribution. The new pipeline has enriched metadata that would give Digital Jay better source awareness.

A) Re-ingest (Replace)

Replace old chunks with enriched versions. Digital Jay gains source attribution: "From Chapter 5 of The Second Act Advantage." Clean, unified schema.

B) Keep Existing

Leave old chunks as-is, add new content alongside. Lower risk but inconsistent metadata across manuscript chunks.

Additional notes (optional)

4. Future Email Content Intake

New content arrives via email from Jay and Skip (AI Tuesday columns, press assets, knowledge additions). How should future content be handled?

A) Automated Pipeline

Set up an email-to-staging pipeline. New content automatically flows into processing. Higher setup cost, hands-free ongoing.

B) Manual for Now

Jonathan manually stages new content when it arrives. Current volume (2-3 items/month) doesn't justify automation. Revisit at 20+ sources/month.

Additional notes (optional)