Below is a detailed lesson plan for **Week 12, Session 1 (Session 23)** of the Prompt Engineering

Specialization course, based on the provided course outline. This session falls under **Part 3: Advanced
Prompt Engineering and Capstone**, with a focus on "Complex Workflows." The topic for this session is
**"Multi-Step Prompting and Orchestration"**, designed to fit a 1.5-2 hour class duration. The plan
integrates lecture, hands-on exercises, and discussion to build students’ skills in creating sequenced
prompt workflows for complex tasks.

---

### Lesson Plan: Week 12, Session 1 (Session 23)

**Topic**: Multi-Step Prompting and Orchestration

**Duration**: 1.5-2 hours (90-120 minutes)

**Date**: Assuming a semester start in late August 2025, this session would occur around mid-
November 2025 (e.g., November 17, 2025), but aligns with the current date context of February 26,
2025, if adapted for a different schedule.

**Location**: Classroom or hybrid (in-person with online access to AI platforms)

**Materials Needed**:

- Access to generative AI platforms (e.g., ChatGPT, Grok, or similar tools via API/browser)

- Laptops for students

- Projector/slides for lecture

- Handout: "Multi-Step Prompting Template" (optional, prepared by instructor)

---

### Session Objectives

By the end of this session, students will be able to:

1. Understand the concept of multi-step prompting and why it’s useful for complex tasks.

2. Design a sequence of prompts to break down a multi-stage problem.

3. Orchestrate prompts to guide an AI model through a cohesive workflow.

4. Evaluate and refine multi-step prompt sequences for efficiency and accuracy.
---

### Session Agenda

#### 1. Warm-Up and Recap (10-15 minutes)

**Objective**: Bridge prior learning and introduce today’s focus.

- **Activity**: Quick review and discussion (5-7 minutes)

- Ask: "Last session, we tested edge cases and adversarial prompts. What’s one limit you discovered
that might affect a multi-step task?"

- Encourage 2-3 students to share insights from their homework.

- **Transition** (5-8 minutes):

- Recap key points from Session 22 (stress-testing, model limits).

- Introduce today’s topic: "Today, we’re moving from single prompts to workflows—how do we guide AI
through multiple steps to solve bigger problems?"

---

#### 2. Lecture: Multi-Step Prompting and Orchestration (25-30 minutes)

**Objective**: Provide a conceptual framework and practical strategies.

- **Content Breakdown**:

1. **What Is Multi-Step Prompting?** (10 minutes)

- Definition: Breaking a task into a series of interconnected prompts where each step builds on the
previous one.

- Examples:

- Writing a story: Step 1 (outline), Step 2 (characters), Step 3 (draft).

- Problem-solving: Step 1 (define problem), Step 2 (list options), Step 3 (evaluate).

- Why it matters: Handles complexity beyond single-prompt capacity.

2. **Orchestration Basics** (10 minutes)

- Definition: Coordinating prompts to ensure smooth transitions and consistent outputs.

- Key principles:
 - Context carryover (referencing prior outputs).

- Clear instructions per step.

- Error-checking between stages.

- Example: "Take the summary from Step 1 and expand it into a detailed plan."

3. **Design Tips and Challenges** (5-10 minutes)

- Tips:

- Start with a high-level plan.

- Use explicit handoffs (e.g., “Use this output for the next step”).

- Test each step independently first.

- Challenges:

- Model memory limits (e.g., losing context).

- Cumulative errors (e.g., Step 1 inaccuracy derails Step 3).

- **Delivery**:

- Slides with a visual flowchart of a multi-step process (e.g., planning a trip: destination → itinerary →
packing list).

- Interactive pause: After orchestration section, ask, “What’s a task you’ve tried that might benefit from
multiple steps?”

---

#### 3. Hands-On Exercise: Building a Multi-Step Workflow (45-50 minutes)

**Objective**: Apply multi-step prompting to a practical scenario.

- **Setup** (5 minutes):

- Divide class into pairs or small groups (2-3 students each).

- Provide a choice of tasks (or let groups propose their own, with approval):

1. Plan a fictional event (e.g., conference, party).

2. Generate a research summary (e.g., topic overview, key points, bibliography).

3. Create a simple game storyline (e.g., setting, characters, plot).

- **Task**: "Craft a Three-Step Workflow" (35-40 minutes)

- Write a prompt to start the process (e.g., “Outline a conference agenda for a tech event”).

- Test it and record the output.

2. **Step 2: Build on the Output** (10-12 minutes)

- Create a second prompt using the first output (e.g., “Take this agenda and list three keynote topics
with descriptions”).

- Test and refine as needed.

3. **Step 3: Finalize the Sequence** (10-12 minutes)

- Write a third prompt to complete the task (e.g., “Based on the topics, draft a 100-word welcome
speech”).

- Test and ensure coherence across steps.

- **Tools**:

- Students use an AI platform to input prompts and save outputs (e.g., copy-paste into a shared doc).

- Optional: Use a simple script (if introduced in Week 4) to automate prompt sequencing.

- **Support**: Instructor circulates to troubleshoot (e.g., “How can you make this handoff clearer?”).

---

#### 4. Group Share and Debrief (15-20 minutes)

**Objective**: Reflect on results and identify improvement areas.

- **Activity**:

- Each group presents their workflow (2-3 minutes per group):

- Share the task and their three prompts.

- Show one key output (e.g., final speech or summary).

- Note one success and one challenge.

- Class discussion (5-10 minutes):

- Ask: “What worked well? Where did the AI stumble? How could you adjust your sequence?”

- Highlight common issues (e.g., context loss, vague transitions).

- **Key Takeaway**:
 - "Multi-step prompting is like directing a play—each prompt is a scene, and orchestration keeps the
story on track."

---

#### 5. Wrap-Up and Homework Assignment (5-10 minutes)

- **Summary** (3-5 minutes):

- Recap: Multi-step prompting breaks down complexity; orchestration ensures flow.

- Preview next session: “Next time, we’ll add multimodal inputs—like text and images—to our
workflows.”

- **Homework** (2-5 minutes):

- Task: “Expand your workflow from today to four steps. Test it, refine it, and write a short paragraph
(100-150 words) on how the extra step improved (or didn’t improve) the outcome.”

- Due: Next session (Week 12, Session 24).

- Submission: Upload to course platform.

---

### Teaching Strategies

- **Scaffolded Learning**: Lecture provides structure; exercise builds skills incrementally.

- **Collaboration**: Small groups encourage teamwork and idea-sharing.

- **Practical Application**: Task options connect to real-world scenarios (e.g., event planning, research).

- **Flexibility**: Adjust exercise time if discussion runs long or students need more guidance.

---

### Assessment

- **Participation (10% of grade)**: Engagement in group work and discussion.

- **Homework (part of 10% Quizzes/Homework)**: Graded for effort, creativity, and analysis (rubric: 1-5
points).
---

### Potential Adjustments

- **If Time Is Short**: Limit presentations to 1-2 groups; others submit written summaries.

- **If Tools Are Limited**: Use one AI platform and provide a pre-set task for consistency.

- **For Advanced Students**: Encourage scripting the workflow (e.g., Python to chain prompts via API).

---

This lesson plan advances students’ skills from single-prompt design to orchestrated workflows, setting
the stage for multimodal integration in the next session and capstone development. Let me know if
you’d like to tweak the task options, add more technical depth, or adjust any section!