CHATBOTS

Nightly Confluence HR Page Indexer for Policy Search

Each night, syncs your Confluence HR space into a Postgres vector index so the policy chatbot always answers from current pages.

CategoryChatbots
Enginesim
Difficultyintermediate
Triggerschedule
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNightly schedule fires
  • ActionList all Confluence HR pages with modified timestampsConfluenceConfluence
  • LogicDiff against indexed versions to find changes and deletions
  • ActionChunk and embed changed pages via OpenAIOpenAI
  • OutputUpsert embeddings and prune deleted pages in PostgresPostgreSQLPostgres

What it does

Keeps the knowledge base behind your HR chatbot fresh. On a schedule it walks every page in the Confluence HR space, chunks the content, generates embeddings with OpenAI, and upserts them into a Postgres table used for retrieval — so answers never go stale after a policy update.

When to use it

Run this alongside any Confluence-grounded HR assistant. Essential when policies change often (comp cycles, benefits open enrollment) and you need same-day accuracy without re-indexing everything by hand.

How it works

  1. 1A nightly schedule kicks off the sync.
  2. 2All pages in the Confluence HR space are listed with their last-modified timestamps.
  3. 3A diff step compares each page against the version recorded in Postgres to find new, changed, and removed pages.
  4. 4Changed page bodies are chunked and embedded via OpenAI.
  5. 5New and updated embeddings are upserted, and rows for deleted pages are removed, in Postgres.
  6. 6A summary of pages added, updated, and deleted is written for the run log.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect ConfluenceSpaces, pages, blueprints.
  2. 2
    Connect OpenAIModels, embeddings, files.
  3. 3
    Connect PostgresAny Postgres URL — query, write, migrate.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.