# Putting Agents Into Real Workflows

**Author** Aklman · **Published** June 8, 2026 · **Language** English
**Canonical** https://aklman.com/sessions/build-with-agents/
**Site** Aklman · aklman.com

> This session is from Aklman. AI reading, quoting, and Q&A are welcome; keep the author attribution and canonical link when quoting.

---

## Summary

A community session on reframing agents from tool lists into work systems made of context, roles, artifacts, and feedback.

## Context

- Format: Community
- Platform: WeChat Group
- Duration: 35 min
- Topics: AI Agents / Workflow Design / Human Agent Collaboration
- Slides: https://aklman.com/sessions/build-with-agents/slides
- Slides PDF: https://aklman.com/sessions/build-with-agents/slides/print
- Print: https://aklman.com/sessions/build-with-agents/print

## Outline

1. **Reframe the agent** — Start from the agent's position in the work, not from models or tools.
2. **Compress work into a schema** — Context, Role, Artifact, and Feedback are the four reusable chunks.
3. **Understand Doer and Tutor** — The same AI behavior can strengthen the work or steal the learning.
4. **Build an explanation frame** — CLT controls input, ICAP deepens processing, and mental models decide transfer.

## Notes

The session moves people from choosing tools to designing repeatable work systems.

The useful unit is not a prompt. It is a loop: context enters, a role acts, an artifact is produced, and feedback becomes the next run.

The long-term value is an inner map: a way to recognize the next situation faster and judge the path with less noise.

## Slides

1. **An agent is not a tool list**
   - 01 / Reframe
   - A better representation is this: an agent is a work unit placed inside context, roles, artifacts, and feedback loops.
   - Tools focus on capability
   - Workflows focus on position
   - Systems focus on feedback
2. **Three primary objects**
   - 02 / Resources
   - A real workflow is not a chat trace. It is a set of objects that can be identified, assigned, checked, and reused.
   - 1 / Context: Goals, sources, constraints, decisions, and current state.
   - 2 / Role: Reader, Planner, Reviewer, and Writer are work boundaries.
   - 3 / Artifact: A document, task, judgment, code change, or reusable record.
3. **How one collaboration flows**
   - 03 / Flow
   - Work is not a single exchange but a loop you can run again: goals go in, judgment comes back.
   - Input / Set the goal: Name what to solve and where the edges are.
   - Context / Gather context: Hand the agent the material, limits, and past decisions.
   - Run / Run and record: Run the tools and produce a checkable artifact.
   - Feedback / Write judgment back: Human trade-offs become the next round's context.
4. **The Doer / Tutor boundary**
   - 04 / Doer vs Tutor
   - The same agent behavior can help you form a mental model, or let you bypass one.
   - Doer / Hands over the answer: It removes search, but also removes understanding.
   - Tutor / Provides scaffolding: It lowers extraneous load while leaving judgment to you.

---

*Aklman · aklman.com · Latest version: https://aklman.com/sessions/build-with-agents/*