Skip to main content

OJT Extraction

Also available in workflows

This specialist runs in chat and as a node you can add to a workflow for repeatable, scheduled runs. See the Agent Catalog for how the two relate.

Overview​

On-the-job training (OJT) plans hold a lot of structure — the project itself and the tasks that make up its curriculum — but usually as prose or tables inside a document. OJT Extraction reads a training document you attach and pulls that structure into clean, ready-to-use tables.

Ellie produces two things: a project summary that captures the metadata for the training program, and a curriculum breakdown with one row per main task, including its duration and detailed description.

Everything comes straight from the document. Ellie extracts only what's explicitly stated, never inferring, and leaves a field blank when the detail is missing. If you ask without attaching anything, it will politely prompt you to add a document.

What you'll need​

How to ask​

Attach the document, then ask Ellie to extract it.

Extract the whole plan:

Extract the project details and curriculum tasks from the attached
training document.

Focus on the curriculum:

Pull just the curriculum tasks from this OJT document, with each task's
duration and description.

What you get​

Two downloadable spreadsheets:

  • Project Summary — the training project's metadata in a single row, such as title, company, approver, location, assignee, dates, and status.
  • Project Tasks — one row per main task, with the task title, duration, and detailed description.

Blank fields mean the detail wasn't in the document.

Tips​

  • Clear section headers and well-formed tables in the source produce the cleanest extraction.
  • Nothing is invented — a blank field means the information wasn't present.
  • Scanned images need to be converted to text first; text-based documents work best.
  • Review both tables against the source before using them downstream.