Course Overview

You do not become fluent in Generative Artificial Intelligence (GenAI) by reading about it. You become fluent by building with it.

What could you do better or faster if you could shape a GenAI tool around the way you actually work?

In this hands-on workshop, participants design and build a custom AI assistant for a real task that matters in their work. Depending on the platform they choose, participants may build a custom GPT, Claude Project, or Gemini Gem.

Over three interactive hours, participants learn how to create assistants that support recurring tasks such as drafting in their voice, serving as a thinking partner for complex work, supporting grant development, or helping teams work more consistently around a shared topic.

Along the way, participants build practical AI fluency by applying what they learn in real time. They also gain a stronger understanding of what GenAI tools do well, where they commonly fail, and how to use them more thoughtfully and effectively.

By the end of the course, participants leave with a working prototype assistant and the confidence to build the next one on their own.

What You Will Learn

Participants will learn how large language models work, including where these tools are strong, where they commonly fail, and why those limitations matter in practice.

The course also introduces practical principles for writing effective instructions and designing more reliable AI-supported workflows. In addition, participants learn how to scope realistic use cases and apply judgment when deciding how AI should and should not be used.

Throughout the workshop, participants apply structured protocols to develop, test, and refine their own custom assistant. As a result, they gain both practical experience and a stronger foundation for future experimentation.

Example Use Cases

Participants build an assistant for a real, recurring task. Common examples include:

• A writing assistant that drafts weekly status updates, memos, social media or monthly summaries in your organization’s standard structure and tone, every time.
• A thinking partner for project ideas that asks Socratic questions, surfaces trade-offs and alternatives, and pressure-tests your reasoning without writing your conclusions for you.
• A grant writing assistant that helps structure specific solicitation response sections—aims, narrative, or work plan—against requirements you provide, while keeping you in the driver’s seat.

Course Format

The course is delivered as one 3-hour instructor-led module, offered virtually or in person. It is organized in two connected parts: Part 1 covers how generative AI works and the principles of effective instruction design, including class discussions, a live demonstration and Q&A; Part 2 is a hands-on build session in which participants design, test, and refine their own assistant, with structured peer feedback and a group debrief. Participants who complete both parts receive a certificate of completion and leave with a working prototype assistant built for one of their own use cases.

Who Should Attend

This course is appropriate for:

• Professionals in any field who want to use AI more reliably and deliberately in their daily work
• Managers and team leads who want to encode their standards into a reusable tool for themselves or their teams
• Teams who want to work more seamlessly across a specific topic
• Knowledge workers who handle recurring writing, analysis, or research tasks—including memos, reports, proposals, and document review
• Funders or fund seekers who want to develop or review grant or funding opportunities more efficiently and effectively.
• Anyone curious about generative AI who wants a practical, hands-on introduction rather than a conceptual overview

Prerequisites

No prior experience with generative AI is required. Participants must have a paid subscription to ChatGPT, Claude, or Gemini to build and test an assistant during the hands-on portion, and should bring a laptop and one real, recurring work task in mind (although we will work together to brainstorm as well!) to build.