Most people making AI ads start with the image. I start with the brief. My background is building systems that can't afford to fail. That same instinct is what makes this work different.
My last role was managing enterprise partnerships — coordinating across academic medical centers, federal institutions, legal teams, and clinical ops simultaneously. That included building marketing collateral with creative teams, developing partner-facing landing pages, and working with PMRC on content that had to be accurate, on-brand, and cleared fast. You learn quickly how to read what a stakeholder actually needs versus what they're asking for, and how to build systems that produce consistent output even when everything else is moving.
I brought that same approach to creative. Every piece in this portfolio starts with a research-backed brief, runs through a structured generation system, and produces output that can be evaluated, iterated, and scaled. That's not how most people approach spec work. It's how I approach everything.
Built a complete, repeatable ad creative infrastructure for a BIPOC-founded Gen-Z skincare brand competing in the hyperpigmentation and ingrown hair categories — from brand spec to 7-agent copy scoring to multi-format ad portfolio.
← Replace placeholder slots with your Topicals ad images
Every generation prompt was built on top of these two reference documents — uploaded alongside each Nano Banana prompt to enforce brand consistency across all outputs.
Grammarly's challenge in 2025–26 isn't awareness — it's "why Grammarly when ChatGPT is free?" Built three ad concepts answering three distinct objections: behavioral anxiety, product proof, and direct competitive positioning.
Olipop's consumer doesn't want to be lectured about gut health. They want something that tastes good and happens to be better. Creative leads with nostalgia as the emotional hook — health stats are the proof layer, not the headline.
DripDrop markets to athletes and performance-seekers. The growth opportunity is everyone else — people who aren't hydrating for sport, but because there's always a next thing. Three underserved personas, three distinct creative angles, each built from real behavioral insight.
"Not for performance. For everything after."
DripDrop's doctor-developed credibility and lower sugar positioning makes them uniquely suited to own the caregiver segment — an audience that doesn't see itself in sports drink ads but is chronically depleted.
Analyzed three real ad accounts from Motion's Creative Benchmarks and Content Engine reports. The goal: identify what actually drives performance, not just what looks good. Two clear principles emerged from the data.
Go face-first. Real person on camera within the first two seconds, every brief going forward. The more interesting test is what happens right after the hook. Eliah's numbers suggest there's conversion left on the table between seconds 4 and 15 — tightening that transition could close the gap between stopping power and actual conversion.
"The first frame is a human handshake, not branding."
Face-first hooks win. Every ad that opened with a real person on camera (Julia: 52.89%, Eliah: 60.7%) cleared 40%+ thumbstop. Text and graphic-led ads (Levi: 13.02%) barely broke 15%. Format is the first decision, not an afterthought.
"A monster hook can carry the whole funnel."
Eliah's hold rate was only 14.75% but still delivered the cheapest conversion in both reports at CA$2.98 per form. Volume of attention at the top compensated for mid-funnel drop-off. Hook quality is the highest leverage variable in the entire ad.
My last role was managing enterprise partnerships in biotech — coordinating across health systems, academic medical centers, legal teams, and clinical ops at the same time. Part of that work was creative: building marketing collateral alongside creative teams, developing custom partner-facing landing pages, and working with PMRC on content that had to move fast without cutting corners. I know what it takes to get something across the finish line inside a regulated, high-stakes environment where everyone has opinions and nothing ships without sign-off.
When I moved into creative strategy, the approach was the same: figure out what the brand actually needs, build a system that produces it consistently, and don't skip the brief. What you're looking at isn't a collection of pretty ads. It's proof I can build the architecture behind creative — the spec, the scoring logic, the prompt system, the production workflow. That includes AI image generation, animation via Veo, and post-production finishing in Canva and Photoshop. The Y2K camcorder treatment on the Topicals animated piece wasn't a happy accident. It was a deliberate creative decision rooted in the brand's actual cultural DNA.
The reason to take a chance on someone who came from clinical ops? I've never once confused pretty output for effective output. Those aren't the same thing — and in performance creative, the difference shows up in your numbers.
All work is speculative. Not affiliated with or commissioned by Topicals, Grammarly, or Olipop.