Classroom Demo

Agentic Workflow Classroom

A guided classroom demo that shows why agents need clear jobs, explicit handoffs, and checks before the final answer.

Demo Request

Plan a 45-minute beginner AI workshop where students learn what agents do, how handoffs work, and when to trust the result.

Shared Classroom Room

Joining room... / Detecting role... / Controls not claimed

Baseline

Start With One Human Request

Workflow Board

From agents to trace

Showing classroom baseline

Clarifier

Turns the vague workshop request into learning goals, audience assumptions, and constraints.

Input
One vague request
Output
Learning goal, audience, constraints

Resource Researcher

Finds examples, materials, and timing limits that fit the clarified goal.

Input
Clarified need
Output
Useful resources, limits, and gaps

Activity Designer

Turns the goal and evidence into a teachable workshop sequence.

Input
Goal plus evidence
Output
Activity sequence and handoffs

Evaluator

Checks whether the plan is supported by the trace and safe to use in class.

Input
Trace plus final answer
Output
Passes, risks, and revisions
Time
Agent Work
Handoff Packet
  1. T1

    Clarifier

    Ask what beginners should understand, what the class already knows, and what constraints matter.

    Learning goal, audience assumptions, constraints, likely confusion, and success criteria.
  2. T2

    Resource Researcher

    Collect examples, materials, time limits, and room constraints that fit the clarified goal.

    Resource list with fit notes, assumptions, source notes, and gaps.
  3. T2

    Activity Designer

    Sketch the learner journey and mark where preparation can happen in parallel.

    Draft workshop flow with activity purpose, timing, and handoff needs.
  4. T3

    Evaluator

    Check the trace for missing inputs, hidden assumptions, weak evidence, and unsafe claims.

    Visible review notes, rejected claims, and a safer workshop plan.

Student Activity

Baseline choice

Students vote together on one improvement. The lecturer uses the group choice to explain what changes in the workflow.

No group choice selected yet.

Trace View

Input Packet

Vague workshop request plus the current learning goal.

Agent Transform

One agent rewrites, researches, designs, or evaluates the packet.

Output Packet

A named artifact that the next agent can inspect and use.