The argument
Three years ago, producing a campaign was the expensive part. Copy, design, variants, localisation, the deck to explain it all: agencies charged for the hours and the hours were real. That cost has mostly gone. A competent person with the right tools can now produce in a day what a team produced in a month.
Two things did not get cheaper. Deciding what to make, and knowing when what you made is wrong. Both still take judgment, taste and a working knowledge of the business, and none of that comes out of a model on the first try, or the tenth.
Most agencies have responded by selling the cheap part at the old price. They call it AI-powered. What it means is the same retainer, the same deck, and a quieter studio.
plexo was set up in 2026 to do the opposite. We treat volume as a given and charge for the judgment: the brief that is actually right, the system that runs without a person babysitting it, the tool built for your process rather than everyone's, and the willingness to say a campaign should be turned off.
Why 2026: this is the year the tools stopped being a demo. Models can hold a brand's rules, draft against them in bulk and check their own work well enough that a small team can run what used to need a large one. That was not true two years ago and it would have been dishonest to sell it.
Why Los Angeles: the clients are here, and so is the creative talent that knows the difference between good and merely fluent. AI raises the floor. It does not raise the ceiling, and the ceiling is what people pay for.