The template was built once. Everything since is dividends.
Fifty presentations later, that decision is the only creative choice I am still grateful for. Not any particular deck, any slide, any argument. The moment I stopped building one-offs and started building a system.
Here is what that system looks like. A single HTML template, pre-wired with analytics, grid and mosaic views, keyboard navigation, full mobile responsiveness, share buttons, a resume-from-last-position feature, and print styles. Built once in the first week, never rebuilt since. Every decision about typography, layout, and transitions made in deck one is still paying dividends in deck fifty. I do not think about those things any more. They are solved. What I think about is the argument.
The nine-step production pipeline converts one idea into four pieces of content across four channels. A presentation. An Open Graph image. A thumbnail. A LinkedIn carousel exported as a PDF. A companion blog post. A 1,200-word X Article. A URL redirect. Mobile optimisation. A claims audit before publication. Nine steps. One conversation with Claude Code. Four separate artefacts. The idea does the work once. The system distributes it everywhere.
This is not efficiency for efficiency's sake. It is a structural change in what a single good idea is worth.
In the old model, one presentation equalled one presentation. A few hours of thinking, a week of production, a slide deck that got presented once and archived. The idea lived and died in a single artefact.
With a pipeline, one clear argument becomes a presentation for people who want to go deep, a carousel for people scrolling LinkedIn, a blog post for people who found it through search, and an X Article for people who want long-form without leaving the platform. Same thinking. Four separate entry points. Each sized and formatted for how that audience prefers to consume.
The infrastructure is the strategy.
The production collapse is real and it is accelerating. A presentation that would have taken a team of three a fortnight in 2020 now takes one person a Saturday. Not because anyone has less talent, but because every element of production: design, writing, formatting, deployment, has been automated or dramatically accelerated. The thinking has not got faster. Everything around the thinking has.
This changes the economics in ways that agencies have not fully reckoned with. When production cost collapses, the billing model built on production time collapses with it. What is left is the part that was always the actual work: the judgement about what to make, the angle that makes someone read past the first paragraph, the decision about what not to build.
The gate is the hardest part.
Every idea that makes it through the pipeline survives a single test: a distinct angle, enough substance for a real argument, and sufficient distance from anything already in the portfolio. Far more than fifty ideas have passed through that gate. Most did not make it. The curation is the hardest part of the entire system. Not the writing, not the design, not the deployment. The decision about what is worth making in the first place.
This is the part nobody discusses when they talk about AI and creative production. The conversation has fixated on output speed, as if the only problem was how long it took to make something. The harder problem is knowing which thing not to make. Better tools lower the cost of a bad idea. They do not raise the floor on what counts as a good one.
What Claude Design changes.
Claude Design launched today. I have not used it yet. From what I can see, it changes the interface from a checklist to a conversation: the nine steps compress into something closer to a single dialogue, and the marginal cost of setting up a new design system falls too. That is not an efficiency gain. It is a category change.
But here is the thing. When production gets easier, the gate gets harder. Every one of the fifty decks is worth making because the editorial bar is high, not because the tools are good. New capabilities that lower friction will produce more content. Most of it will be undifferentiated. The system advantage is not speed. It is the judgement layer that decides what goes through.
The compound returns in creative work are not in making things faster. They are in building the infrastructure that makes every good idea go further.
That is the thing worth building. And the thing worth protecting, regardless of whatever tools arrive next.