Creative Engineering: A Manifesto for the AI Era
Every technology wave separates the people who make things from the people who make the systems that make things. Printing separated the writer from the printer. Television separated the storyteller from the broadcaster. Software separated the marketer from the engineer, and for thirty years that separation has defined how marketing gets done. Creative people had ideas. Engineers had systems. A whole industry grew up in the middle, translating one side to the other and charging for every handoff.
AI closes the gap. The machine does not have taste. What has changed is the medium. The output of a marketing team in 2026 is no longer a campaign. It is a system that produces campaigns: a model that knows the brand, a workflow that turns a brief into drafts, a set of rules that decides what ships and what gets a second look. Once the medium is software, the person with the idea and the person who builds the system have to be the same person, or at least sit at the same desk.
We call that person a creative engineer.
A creative engineer thinks about a brand the way an engineer thinks about a codebase. Where does the voice live? Who can change it? What happens downstream when it changes? Most companies cannot answer these questions. The voice lives in a PDF, three people can change it, and when it changes nobody tells the AI tools the team uses every day. The result is familiar. Every draft starts off brand and gets fixed by hand, which is the most expensive way there is to be consistent.
So, a few things we believe.
The deliverable is a system, not a project. A project ends. A system runs. If a piece of work does not leave behind something the team can operate without us, we did it wrong.
Taste has to be encoded. A brand that lives only in people's heads disappears at every handoff. Voice, visual standards, messaging, and the judgment calls behind them can all be written down in a form a machine can apply. Written down, the craft travels.
Build where the team already works. The best system is the one nobody has to remember to open. We build inside the AI platform a company already pays for, because a second tool is a second place for the brand to drift.
Transparency is a discipline. Every hour, every decision, and every rule the system applies should be visible to the client. Opaque work is where margin hides.
The team operates the system. The client's people run their marketing, on software we built with them. That is the difference between a vendor and an engineer.
Why now? Three things happened at once. Models got good enough to hold a brand's voice across thousands of outputs. Companies started paying for AI platforms without a plan for what to put in them. And the cost of building bespoke software fell far enough that a marketing team can have its own. Any one of these is interesting. Together they mean the old separation is finished.
The interesting question is who does the engineering: the vendors who sell the tools, the platforms that own the models, or the people who understand the brand. We think it should be the last group, and we think they need help.
That is what Retent is for. A new era of work begins now.
All articles