AI Prompt Engineering for Reliable Results

This advanced workshop is for engineers, test directors, and technical staff who know AI tools are supposed to help but haven't yet found consistent, practical ways to use them in their daily work. Instead of generic advice about “being polite to the chatbot,” participants learn how to give AI the right context, use an interview-style approach to clarify tasks, turn repeat documents into reusable templates, and evaluate output with clear, objective criteria. These skills are practiced using the real documents and workflows they already use every day. The course is tool-agnostic, so the techniques transfer across Gemini, Claude, and Copilot, and it follows a bottom-up model that does not require enterprise integration. Every workflow includes a human review step, and the program ends with hands-on labs that produce useful, organization-specific materials participants can take back to work.

Part 1 — On-Site Kickoff (3.5 Hours) 

The onsite kickoff training session will include 5 lessons:

1. From Grocery Store to Recipe — a practical introduction to where AI fits into technical work, what your approved tools can and can’t do, and how to use them to support your work rather than replace your judgment. We’ll also cover the boundaries for handling data and explore common use cases by role. 
2. Context Is the Prompt — learn how to give AI the background it needs, including your role, responsibilities, and organizational context, so the output is more relevant and useful. Lab: build a reusable prompt that captures your role and team context. 
3. The Interview Pattern & Delegation — practice using AI as a thought partner that asks the right follow-up questions before jumping to an answer. This helps you define the task clearly and get better results from the start. Lab: work through a real task from initial idea to a well-scoped request. 
4. Advanced Prompt Mechanics — understand the prompt-writing techniques that improve results, including how information order affects output, how to make requests clear and testable, and why AI can seem to “forget” important details. Lab: revise a weak prompt into one that is clear, reliable, and easier to evaluate. 
5. Hands-On Lab & Use-Case Backlog — put the techniques into practice using one of your own documents or workflows. You’ll leave with a completed example you can use right away, along with a shared list of future use cases to guide the follow-on sessions. 

Part 2 — Optional Follow-On Sessions

Additional follow-on sessions can be scheduled to get more in-depth coverage of essential topics. 

Session A: Document Templating, Bottom-Up (3 Hours) 

This flagship session shows participants how to turn a document they create over and over again such as a report, summary, test artifact, or project update into a reusable prompt-driven template. The focus is on starting with real work, not abstract theory: identifying the repeatable sections, deciding what information should stay fixed versus what changes each time, and building a practical prompt structure that helps generate a strong first draft. Participants leave with one working template they can reuse immediately, along with a repeatable method for creating more templates across their team. • 

Session B: Defining "Done" — Rubrics & Evaluation (3 Hours) 

Participants learn how to move beyond “this looks pretty good” and evaluate AI output with clear, shared standards. This session introduces simple scoring criteria for accuracy, completeness, tone, structure, and usability so teams can decide whether an output is ready to use, needs revision, or should be discarded. It also covers why evaluation matters when models change over time, and how a lightweight rubric can help teams compare outputs more consistently across tools, prompts, and future model upgrades. 

Session C: Master Prompts & Shared Agents (3 Hours) 

This session helps teams turn individual prompt successes into shared assets others can actually use. Participants learn how to organize a strong master prompt, separate reusable instructions from task-specific inputs, and document enough context so a teammate can get reliable results without starting from scratch. The session also explores how shared prompts and lightweight agents can support common team workflows while still preserving review, control, and accountability. • 

Session D: From Test Plan to Test Report (3 Hours) 

Using a complete technical workflow, this session shows how prompt engineering can support work from beginning to end rather than as a one-off shortcut. Participants walk through how to structure prompts for planning, drafting, revising, and summarizing across the lifecycle of a test-related deliverable. Along the way, they apply templating and evaluation techniques to keep outputs aligned with the intended purpose, audience, and standard of quality. The result is a practical example of how AI can support a repeatable workflow without replacing technical judgment. 

Session E: Requirements & Document Analysis (3 Hours)  

This session focuses on one of the most useful technical applications of AI: helping people make sense of dense, overlapping, or inconsistent documents. Participants learn how to prompt AI to compare requirements, identify possible conflicts or duplication, surface missing information, and summarize large documents in a way that is actually usable. Emphasis is placed on verification and human review, so the output becomes a faster starting point for analysis, not a substitute for responsible decision making. 

Session F: Multimodal & Generative Output (3 Hours) 

Participants explore how to work with more than plain text, including PDFs, Office documents, images, and other mixed-format inputs that are common in technical environments. The session covers how to ask for the kind of output you actually need whether that is a summary, extraction, transformation, rewrite, or structured deliverable and how to give enough guidance so the format is useful the first time. It also addresses the limits of multimodal tools and where human review is especially important. 

Session G: Process & Day-to-Day Productivity (3 Hours) 

This session is designed for the small, repeatable tasks that quietly consume time every week. Participants learn how to use AI to support meeting recaps, transcript-to-summary workflows, roll-up reports, recurring communications, and other routine knowledge-work tasks without losing clarity or control. The emphasis is on building simple, reliable workflows that reduce manual effort while still fitting into the team’s normal review and approval process. 

NOTE: Data handling & safety boundary is taught up front and reinforced every session: approved tools only, sanitized/unclassified material, metadata stripped before upload, and a human review before any output is reused.

Each student receives a comprehensive set of materials, including course notes and all class examples.