DOCUMENT OPS

Route low-confidence contract classifications to a Slack human-review gate before filing

Classifies each incoming Dropbox contract and, when the model's confidence is below threshold, posts it to a Slack review channel for a human to pick the correct type before it…

CategoryDocument Ops
Enginesim
Difficultyintermediate
Triggerevent
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew contract in Dropbox intake folderDropboxDropbox
  • ActionExtract text and classify with confidence scoreOpenAI
  • LogicBranch on classification confidence threshold
  • ActionPost low-confidence cases to Slack review channelSlack
  • OutputCreate Notion row using confirmed typeNotionNotion

What it does

Adds a human-in-the-loop checkpoint to contract filing. High-confidence classifications file themselves into Notion; ambiguous ones pause and ask a reviewer in Slack to confirm the contract type before anything is written.

When to use it

Use this when misfiling a contract is costly and your folder receives oddball or hybrid documents that an LLM alone gets wrong. It keeps automation fast for the easy 80% while protecting the tricky 20%.

How it works

  1. 1A new contract in the Dropbox intake folder triggers the run.
  2. 2The PDF text is extracted and sent to OpenAI, which returns a contract type plus a confidence score.
  3. 3A logic branch checks confidence against your threshold.
  4. 4If confident, the row is created in the correct Notion register immediately.
  5. 5If not, a Slack message with the suggested type and document summary is posted to a review channel with type-choice buttons.
  6. 6The reviewer's selection determines the Notion register, and the row is created with the human-confirmed type recorded.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect DropboxFiles and folders.
  2. 2
    Connect OpenAIModels, embeddings, files.
  3. 3
    Connect SlackChannels, DMs, threads, mentions.
  4. 4
    Connect NotionPages, databases, comments.
  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.