AI made the work faster - it didn't shrink the retainer

A prospective client asked us last month, politely, whether our retainer should shrink now that we're "obviously" using AI in the process - twenty hours a month instead of the forty or more a growing brand usually needs. It was a fair question. It's also built on a false equivalence, and we think most studios are about to hear a version of it whether they're ready for it or not.

When speed reads as a shorter contract

The instinct comes from a real shift. According to a 2025 survey by Productive, an agency-management platform, of more than 180 creative and digital agencies, pricing was the most divided question in the entire study: roughly a third of agencies had already fielded an explicit "AI discount" request from a client, with close to half expecting one soon. Agencies that sell a raw hour are right to worry - a faster hour genuinely is worth less when the hour itself is the product.

Our retainer is structured around a monthly allocation of time, but the value has never been limited to the hours alone. That structure gives the relationship more stability than pure hourly billing, but it does not make a client's underlying question disappear - if the work got faster, why hasn't the number moved.

What actually fills the hour now

AI didn't take work out of the hour. It took out the slowest, least interesting part of it - producing the raw variations - and left the hour fuller with the part that was always the actual job. AI may help us produce initial directions faster, but it does not understand a cybersecurity product, recognize when a visual quietly undermines trust, or know which idea will hold up consistently across a website, a pitch deck, a campaign, and a product interface. A set of logo directions that used to take a day to sketch now takes an afternoon to generate. The day we used to spend sketching doesn't disappear from the invoice; it moves to that judgment call - telling which of twelve generated options respects three years of brand decisions nobody wrote down, which one is one edit from right, and which nine were never going to work no matter how fast they arrived. That review-and-edit pass still has to happen by hand, at the same pace it always did, before anything is fit to publish under a client's name.

Same twenty, forty, or eighty hours, then, now produce more finished thinking inside them - more directions actually tested, more rounds reviewed, faster turnaround on the same brief. AI is one more tool in that hour, not a reason to book fewer of them. It raises what the hour is worth; it doesn't shrink how many a brand needs.

The hour was never priced on typing speed

Strip away the generation step and most of what a retainer hour buys was never about how fast anyone can produce a variant. It's priority when something's urgent, brand context that doesn't need re-briefing every month, and the judgment to say no to an on-brief idea that's quietly off-brand. None of that got faster this year, and none of it is what a client is really objecting to when they ask for a shorter retainer - they're reacting to faster execution, which is exactly the one part of the hour that did change.

The question is no longer how many hours a design took. It's whether the studio helped the client make a better decision, protect the brand, and move faster with confidence.