Facilitator Presentation Practice Coach

This AI-powered Facilitator Presentation Practice Coach was designed as a prototype learning tool to support workplace presenters preparing for live workshops and internal presentations. Through guided voice rehearsal and structured feedback, facilitators can practice delivery in a low-stakes environment that promotes confidence, clarity, and readiness before stepping in front of an audience. 

• Audience: Workplace facilitators, Subject Matter Experts (SMEs), and new managers preparing to lead presentations or workshops 

• Responsibilities: Instructional design, conversational flow design, AI prompt engineering, voice interaction protocol development, feedback framework creation, iterative testing and refinement 

• Tools Used: ChatGPT (Custom GPT Builder) 

Project Overview

The Facilitator Presentation Practice Coach is a custom GPT prototype designed to support workplace presenters preparing for live workshops and internal presentations.

This tool provides a low-stakes rehearsal environment where facilitators can practice a short speech aloud and receive structured, transcript-based coaching focused exclusively on delivery.

The goal of this prototype was to explore how AI can be used as a scalable, on-demand practice partner for employees who do not always have access to live coaching.

The Challenge

Many Subject Matter Experts (SMEs) and facilitators are highly knowledgeable in their content but may struggle with presentation delivery. Organizations often rely on post-session feedback, which arrives too late to meaningfully improve performance. While live coaching can be effective, it is not always scalable or accessible for teams with limited time or resources.

I began exploring how a tool built with technology many organizations already have access to—such as ChatGPT—could provide a more practical alternative. Instead of requiring a multi-hour eLearning course or instructor-led training session, this solution offers a cost-effective, on-demand way for learners to practice their delivery whenever it is convenient for them.

This prototype explores how a voice-enabled AI tool could provide formative practice and coaching before a live presentation.

Design Decisions

To create an effective rehearsal experience, I intentionally structured how users interact with the tool and how feedback is delivered. These design decisions were informed both by instructional design principles and by the technical constraints of voice interactions within custom GPTs. 

Feedback was deliberately directed toward speaking and facilitation skills rather than the content of the presentation. Because language models tend to default to evaluating content, I had to carefully guide the model to focus on delivery skills such as confidence, pacing, and clarity—the areas this coach is intended to support. 

Through iterative prototyping and refinement, I shaped the tool to meet its intended purpose. The following design elements illustrate how the experience was structured to support clear guidance, focused practice, and actionable coaching. 

Takeaways: Platform Constraints & Design Considerations 

Designing within ChatGPT’s voice environment required thoughtful adaptation and multiple rounds of testing and refinement. Rather than treating platform limitations as barriers, I approached them as design constraints that informed how the experience was structured.

Transcript-Based Analysis

Although the tool uses voice interaction, custom GPTs do not have access to direct audio metrics such as vocal volume, pitch range, tone variation, or precise speech rate. As a result, feedback is based on transcript cues and linguistic patterns rather than acoustic analysis. I designed the feedback framework accordingly so the coach focuses on delivery signals that can be inferred from speech structure, phrasing, and pacing.

Voice Interface Limitations

The voice interface also introduces several interaction constraints. Custom GPTs cannot detect when voice mode is activated, cannot automatically initiate speech without user input, and may not consistently display transcripts after a voice session ends.

To address these constraints, I designed the experience to include:

• A clear voice initiation flow for starting practice

• Recovery language to handle interruptions or pauses during a speech

• A structured feedback approach that reinforces learning even during short practice attempts

User feedback during testing highlighted the value of having an accessible, repeatable practice tool. Kathryn Trites, Graphic, UX, & UI Designer, had this to say: 

"This is so cool! My impression is that it's a great starting tool. I could see myself using this if I needed to practice presenting something. The fact that it provides feedback is so helpful and that you have unlimited control over the number of times you can record and evaluate yourself is also great." 

🪄See the Facilitator Presentation Practice Coach in action in the video below. 

🪄Try the Facilitator Presentation Practice Coach using the script below.

Sample Practice Speech

If you'd like to try the Facilitator Presentation Practice Coach yourself, but can't think of what to say, you can practice using the short speech below. This sample script represents a typical workplace update and is designed to take about 30 seconds to deliver. This script can be used to test the voice coaching experience in the custom GPT.

Tip: When using the GPT, read and follow the initial text guidance from the GPT.