Moving beyond curiosity into high-velocity production pipelines. How I use custom AI agents, automated moodboarding, and generative tooling to collapse briefing times and scale output.
I do not use AI to generate final logos or packaging illustrations in one shot — because automated generic output lacks brand soul and retail nuance. I use AI to automate the administrative friction of commercial design: summarizing meeting transcripts into structured briefs, generating 40 packaging concept directions in 30 minutes, and accelerating handoff so 90% of my energy stays on craft.
Select a production stage to simulate how custom agents and generative models collapse traditional turnaround times across client projects.
Pulls unstructured Slack notes, PDF meeting transcripts, and retail guidelines into an explicit typographic and structural brief.
Built custom AI agents (like the "Ask Shiv's AI" widget on this portfolio) that can handle customer queries, answer design questions, and pull from a knowledge base — things that used to eat up time I should have spent creating.
Took the AI tools off the mood board and put them into actual production pipelines. Automated parts of the code generation, testing, and layout process — which sounds less exciting than it is, but it's what separates a working system from a prototype.
Used AI for voice work, video editing, and early concept generation. The biggest win was using generative tools to build rough visual references and packaging mocks before opening Illustrator — it cut briefing time from days to a few hours, and it meant the team was aligning on a direction, not still debating one.
Back in college when generative AI was just becoming something people could actually use, I was already running workshops — showing other students what these tools could do for art and design work. It was early, a bit experimental, but that's how I got comfortable with the space.