AI & RAG

Sales-Engineer Answer Bot Grounded in Loom Walkthroughs and Docs

Watches a Slack channel for customer technical questions, retrieves the most relevant Loom walkthrough transcripts and Confluence docs.

CategoryAI & RAG
Enginesim
Difficultyintermediate
Triggerevent
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew question posted in #se-help Slack channelSlack
  • LogicClassify message as answerable technical question
  • ActionRetrieve matching Loom transcripts and Confluence docsLoomLoom
  • ActionDraft grounded answer with citations via OpenAIOpenAI
  • OutputPost cited draft reply in Slack threadSlack

What it does

When a teammate posts a technical question in your #se-help Slack channel, this workflow finds the matching answer across your recorded Loom walkthroughs and Confluence runbooks, then drafts a grounded reply with source links so the sales engineer never reinvents an answer they have already explained on camera.

When to use it

Use it when your SE team keeps fielding the same integration, config, and "how does feature X behave" questions in Slack, and the real answers already live in past demo recordings and docs nobody searches.

How it works

  1. 1A new message in the watched Slack channel triggers the flow.
  2. 2A logic step classifies whether the message is an answerable technical question; non-questions stop here.
  3. 3The retriever embeds the question and searches indexed Loom transcripts plus Confluence pages for the closest passages.
  4. 4OpenAI composes a concise answer strictly from the retrieved passages and lists each source.
  5. 5The draft, with Loom timestamps and doc links, is posted back in-thread for the SE to review and send.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect SlackChannels, DMs, threads, mentions.
  2. 2
    Connect LoomVideo transcripts, libraries.
  3. 3
    Connect ConfluenceSpaces, pages, blueprints.
  4. 4
    Connect OpenAIModels, embeddings, files.
  5. 5
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  6. 6
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  7. 7
    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.