make digital instruction
more approachable.
Meet Cynthia
- Project
- Project Type
- Project Status
- My Role
- Tools
- Focus
- Blog Post
- Cynthia: AI-Assisted Microsoft Teams Tutorial
- Self-Directed AI-Assisted Instructional Video
- Completed Self-Directed Project, 2025
- Concept Developer, Content Strategist, Visual Designer and Video Producer
- Midjourney, ElevenLabs, HeyGen Avatar IV, OBS Studio and CapCut
- AI Avatars, Instructional Content, Screen Recording and Video Production
- Read the Full Project Walkthrough
Project Details
Cynthia began as an exploration of how AI-assisted presenters could make software tutorials feel more natural and approachable than traditional PowerPoint avatars and cartoon animations. I wanted to test a production method that did not require actors, a studio, or a large production team.
For the proof of concept, I created a short tutorial showing how to change a background in Microsoft Teams when the user is not already inside a scheduled meeting. I selected the topic because it addresses a small but common workplace question that often comes up when someone is preparing for a call.
I developed the concept and instructional message, wrote the script, designed Cynthia, recorded the Microsoft Teams demonstration, and produced the final video. This was a self-directed experiment rather than official Microsoft or employer training, allowing me to evaluate the workflow while retaining control over the presentation, pacing, and final edit.
Production Workflow
I combined AI-assisted tools with screen recording and traditional video editing to create the tutorial. Each platform served a specific purpose while I directed the instructional message, visual presentation, timing, and final edit.
1. Concept and Script:
I selected a common Microsoft Teams question and wrote a concise script that divided the process into clear, manageable instructions.
2. Avatar Design:
I used Midjourney to create Cynthia as an approachable and professional presenter suited to workplace instruction.
3. Narration and Animation:
I generated the narration with ElevenLabs and used HeyGen Avatar IV to animate Cynthia and coordinate her delivery with the voice track.
4. Screen Demonstration:
I recorded the Microsoft Teams process with OBS Studio, following the same sequence presented in the narration.
5. Editing and Delivery:
I assembled the avatar footage, narration, screen recording, captions, and transitions in CapCut, adjusting the timing so the presenter and software demonstration worked together.
Challenges and Creative Decisions
Because several parts of this workflow were new to me, the project required testing, comparison, and adjustment before I reached the final production approach.
Selecting the Animation Tool: I initially tested Kling AI for Cynthia’s animation and lip-sync, but the mouth movements did not feel natural enough for the tutorial. I moved the character into HeyGen Avatar IV, which produced a smoother result and worked directly with the ElevenLabs narration.
Coordinating Instruction and Demonstration: The narration, avatar footage, and Microsoft Teams recording needed to remain synchronized. I adjusted the timing so each action appeared when it was being explained rather than asking the viewer to connect separate pieces of information.
Choosing the Editing Workflow: CapCut provided an efficient way to assemble and refine this short prototype. The experience also helped me determine that Adobe After Effects or Premiere would offer greater control for future tutorials requiring more complex animation or detailed visual guidance.
What I Took From It
This project reinforced that a more human-looking presenter does not automatically make instruction more effective. The avatar, narration, and movement still need to support the task without pulling attention away from the software demonstration.
It also strengthened my approach to evaluating emerging tools. I learned to compare outputs, change direction when a result was not strong enough, and select production methods based on instructional clarity rather than novelty or speed alone. The workflow now gives me a foundation for future AI-assisted tutorials while leaving room for more advanced editing when the content requires it.
Read the full project walkthrough for a closer look at the tool testing, production decisions, and ideas I identified for future tutorials. You can also return to the collection to explore more of my AI-assisted creative work.