Explore a task
Start with work you recognize.
Follow what goes in, what the system does, and what comes out. These examples are scripted teaching cases; no model calls or external actions run here.
Everyday
Answer a question about your documents
- Input
- A question and a small set of documents.
- What happens
- Retrieve relevant passages, then draft an answer tied to that evidence.
- Output
- An answer with source references and visible gaps.
See why relevant evidence matters more than a confident answer.
Follow this example →Business
Prepare a weekly project report
- Input
- Previous reports and current project updates.
- What happens
- Collect evidence, reconcile changes, flag missing updates, and draft for review.
- Output
- A report draft with sources and unresolved items.
See where a fixed workflow helps and where a person decides.
Follow this example →Engineering
Generate a new test project
- Input
- A DUT brief, a past project, and a shared Python framework.
- What happens
- An agent creates project files using existing capabilities within enforced boundaries.
- Output
- Reviewable code, configuration, and a record of checks and gaps.
See how a harness surrounds a model with tools, context, and permissions.
Follow this example →Teams of Agents
Have an agent team improve your brief
- Input
- Your goal, representative files, and intended human role.
- What happens
- A lead coordinates a writer, parallel receiving agents, and independent reviewers.
- Output
- A revised brief, candidate approaches, and decisions that still need you.
See delegation, disagreement, and bounded revision in one application.
Follow this example →