Introduction to Generative AI: Prompt Engineering and Tools Overview

Sample content placeholder

Situation

After several months of access to Microsoft Copilot licenses, many associates weren’t fully adopting the tool. They defaulted to manual workflows, limiting productivity gains, and struggled to see how Copilot could act as more than just an automation tool. The gap wasn’t access but confidence and practical understanding. To address this, I partnered with the CIO to create a training program that demonstrated how Copilot could streamline everyday tasks and also support higher-level, strategic work.

Confidence distribution (Pre vs Post)

Shares of learners at low / moderate / high confidence levels before and after training.

Solution

I designed a personalized, modular e-learning program released on a drip-feed schedule. The training began with a module on AI foundations and company policy, where learners could test out or complete a recap. From there, associates learned core prompt engineering skills, followed by targeted Copilot training across Teams, Outlook, Word, PowerPoint, and Excel. Each module offered test-out options for experienced users while guiding others step by step.

The final module highlighted how advanced users leverage Copilot as a strategic partner, supported by embedded resources and external links to keep the program agile as the tool evolved.

Process

The training was developed using the SAM model, combining discovery, prototyping, and iterative development. A detailed scope document and SME interviews informed the design.

The program was developed using an alpha–beta–gold cycle to refine flow and accuracy. Evaluation was built in from the start, using Kirkpatrick Levels 1–3, pre- and post-surveys to measure confidence gains, IT license data, innovation tracking, and a final readiness assessment.

Tools Used

Articulate Storyline
Microsoft Copilot
Microsoft Word
Microsoft Visio

Outcome

Toggle between total user adoption and month-over-month percent growth.

Learners reported a 29% increase in confidence using Copilot, and IT data showed an 18% lift in license adoption. A final assessment validated knowledge retention, with 86% passing on the first attempt. Associates also submitted dozens of AI-driven ideas, several of which informed new business solutions. Feedback highlighted that the modular, scenario-driven approach was more engaging than traditional systems training.

Challenges

Copilot’s frequent updates risked making static content obsolete. To address this, I embedded videos directly from Microsoft and YouTube so they could be swapped quickly as the tool changed. Another challenge was balancing advanced and novice learners; I solved this by creating personalized learning paths with test-out options and optional deep dives. To avoid scope creep, I narrowed the program to core concepts while pointing advanced learners toward curated external resources.

Lessons Learned

This project reinforced the need to keep training agile for rapidly changing tools. It also showed the value of simplifying to core concepts, while offering personalization for diverse audiences. Finally, I saw firsthand that learners engage most when training connects to real problems, not just features.