This interactive video lets learners practice difficult conversations in a realistic, branching format with instant feedback, so they can see the impact of their choices and better transfer those skills on the job.
Audience: Managers, Team members
Focus: AI-powered interactive decision-making video
Responsibilities: Instructional Design, Generative AI, Prompt Engineering, Script writing, Video editing, Visual design
Tools Used: Synthesia, Invideo AI, Miro, Adobe Premiere Pro, Chat GPT, Canva
This project was created in 2025.
Unfortunately, the interactive video is restricted to my work email, so I have made a video recording to demonstrate the interaction.
Talking-head videos make people watch, not try. There’s no choice, no feedback, and the content feels generic—so attention drops and skills don’t stick.
I used Synthesia’s interactive features to turn a static video into a branching, choose-your-reply scenario with instant feedback. Learners make decisions, see the consequences in-scene, and leave with practical next steps they can apply immediately—making the video relevant, engaging, and memorable.
To build this course, I first created a short AI-generated video in Invideo to set the scene, applying Gagné’s “gain attention” and storytelling. I used ChatGPT to outline the flow, draft dialogue, and craft plausible A/B choices with immediate feedback (Gagné’s present content → elicit performance → provide feedback). I mapped the branching in Miro as a concise storyboard. Finally, I produced the interactive video in Synthesia, leveraging branching for real-time responses
I used Invideo with a short generative-AI prompt to produce a 10-second opener that sets the stage. I iterated by giving the tool new instructions until the visuals and timing matched the moment I wanted. Then I pulled the clip into Premiere Pro to add the notification button detail I needed. The clip exists purely to grab attention per Gagné’s “gain attention”—a quick hook, no solutions, just context so learners lean in.
I started with the prompt: Create a 5-10 second video. Start with people in an office packing up and then pan over to the computer where the clock shows 5pm. Then have a message ping on the computer screen.
After I had the initial draft, I then was able to further edit each individual scene (image to left)
With ChatGPT, I structured a branching dialogue with paired, plausible choices, drafting a text storyboard on Miro. This supports deliberate practice and choice architecture: learners act, see a reaction, and immediately receive concise feedback, aligning to Gagné’s “elicit performance → provide feedback.” Generative AI sped up drafting; I refined tone, simplified jargon, and ensured every screen advances one behavior. I then built the video in Synthesia, making sure that all of the interactive buttons are connected to the right scenes.
Using an interactive, branching video (vs. a talking-head clip) proved more engaging and effective because learners make choices and see immediate consequences.
Overall, this format felt more relevant and hands-on than a static video, and it’s easy to maintain in Synthesia since it’s cloud-based. The one drawback was that a few avatar reactions didn’t match the moment (e.g., smiling on a “negative” branch). Even so, it delivered realistic practice with clear carryover to on-the-job conversations.