HUMAN interfaces
The first element of intuitive design.
E.A. Rockett — AI Field CTO & Designer · Advisor to Boards & C-Suite
The layer between people and AI. Built for creative enterprises first: film, media, music, advertising. And anywhere the work is the product.
Everyone has access to nearly the same foundation models now. So the model is not what separates the companies getting real value from the ones still running pilots. What separates them is the layer between that technology and the person doing the work.
A human interface works when three things are true: it's intuitive, it works, and the output can be trusted and used. Miss any one and people quietly go back to the old way.
A human interface that works relies on knowing and effectively using the technology and data underneath it: the tools your teams already use, the AI models you deliberately chose, and the ghost AI that arrived inside products you bought for other reasons. Nobody chose it. Nobody owns it. It's in the work daily.
Three services span the whole arc: decide what's worth building (the strategy), design and build it (the human interface), then make it something people actually reach for and tell others about (the marketing).
Where AI actually stands in the business: the deliberate AI model choices, and all the ghost AI nobody selected. What to bet on, and the training-data, licensing, authorship, and provenance questions that decide whether any of it can ship. The more you produce, the more this matters: every generated asset carries a question about what the model was trained on and who owns the result. I know that ground from several sides: co-founder of SPDX, the open-source provenance standard under the Linux Foundation; patent examiner and trademark examining attorney at the USPTO; and lead copyright and licensing counsel for Motorola's global software portfolio.
I design and build the layer your people actually touch. It sits on top of everything else: your existing tools and data, the AI you deliberately chose, and the ghost AI you didn't. I build the prototypes myself, so decisions get made against something running rather than something described. Underneath, whatever the job needs: agents, retrieval, workflow integration, or none of the above, when the right answer is software with no AI in it. An expert already moves at the speed of their own thinking. The test is whether the interface keeps up, or breaks the pace. I started as a UI/UX designer. Same problem ever since; the technology underneath just changed.
It's a go-to-market problem, and it works the same whether the audience is your own teams or your customers: there's a product, an audience that hasn't bought in, and a value proposition nobody should have to explain, because the human interface is so obvious it already did. That's the rare part. I run both the same way: positioning, narrative, enablement, champions, measurement. Trust is the part most launches skip. People adopt human interfaces that are intuitive, do what's expected, and can be trusted. Push too hard and they resist.
Co-owned product strategy for Adobe Firefly Generate Soundtrack, Adobe's generative AI music product, from personally built prototype through public release. Worked directly with ML researchers on model behavior and guardrails. Built an AI document translation tool that spread from the legal department to the whole corporation.
A no-code configuration layer that let attorneys change contract language without developers. A self-serve interface over 1.5M+ pages of contracts for non-technical business users. An analytics layer inside a self-publishing platform, built for authors with no analyst support.
Co-led AI@Adobe, which brought third-party AI tools to roughly 30,000 employees. Recognized in 2025 by Great Place to Work® for innovation leadership. Designed the Customer Zero validation program across products serving 850M+ monthly active users; its findings moved published launch dates.
Led exploratory training-data licensing discussions with major record labels and publishers, translating rights-holder economics into workable commercial structure. Co-founded SPDX under the Linux Foundation.
Led NOOK's shift from device-centric to digital-first: a $450M+ improvement in annual financial performance. Also co-led the Motorola–Google–Verizon collaboration that launched Android and formed the Open Handset Alliance.
Named alongside Howard Behar (Starbucks) and Indra Nooyi (PepsiCo) as a “unicorn leader” by Phil Le-Brun, Executive in Residence at AWS, and a credited contributor to his book The Octopus Organization: A Guide to Thriving in a World of Continuous Transformation (Le-Brun & Werner, 2025). Keynotes and panels at AWS re:Invent, CxO Institute (New York, Chicago, San Francisco), GDS CIO Summit, Dreamforce, and The Future of Music Conference; podcasts with Le-Brun on AWS Executive Insights and with Jon Herstein, Chief Customer Officer of Box, on AI-First.
With Jon Herstein, Chief Customer Officer of Box. On putting AI in front of people who have to use it, and what changes when they do. 2026.
With Phil Le-Brun of AWS. On saying yes to the work rather than governing it into stillness. May 2025.
Panel on lessons from the frontlines of implementation. Las Vegas, December 2024.