AI
Scroll to explore the AI-assisted projects and the workflow behind them below.
NewHolme
AI workflow · Motion · Creative experimentation
NewHolme was where I first began exploring AI video as part of a real client workflow. Creating these projects gave me the opportunity to experiment with prompting, image generation, voice synthesis and editing, gradually refining a process that would continue to evolve with every new project.
AI presenters started as talking-head videos videos which could be made without filming, reducing production time and costs while giving clients more flexibility.
At the time, combining AI-generated scenes with carefully selected stock footage was often the best approach, and I enjoyed experimenting with different workflows to understand what the technology could realistically achieve. They also show how quickly both the technology and my workflow evolved.
This is a simplified example of the workflow for one of the videos:
Ascentis
Fully AI-generated · Workflow optimisation · Video generation
Only a few months later, my workflow had already evolved again. Working on a monthly retainer for Ascentis gave me the opportunity to create fully AI-generated videos, no longer relying on stock footage as the technology continued to improve.
These days, my workflow centers around Kling, HeyGen and Nano Banana, but I’m always testing new tools as they emerge.
AI B-roll
Agency showreel · Multi-client · AI b-roll
Beyond presenter videos, I used AI to create bespoke b-roll for our agency showreel, bringing together work from multiple clients into a single visual story. It was a chance to experiment with placing brands into believable environments, making digital projects feel more tangible and engaging than traditional mockups alone.
The final video combine AI-generated footage with traditional motion graphics where appropriate. Almost all presenter-led scenes and human interactions were created using AI, while graphic animation was used to support the overall story.