A salon owner in a hurry opens an AI tool, types a prompt for an Instagram ad, and gets back something clean in under a minute: nice typography, a tidy layout, a stock photo of a woman with glossy hair. It looks like an ad. It could be an ad for any salon in any city. The owner posts it anyway, because it took one minute instead of two hours, and the clock felt like the thing that mattered.
That trade, speed for sameness, is the real question AI raises for small business advertising. Not whether the tools work. They do, mostly. The question is whether moving faster is actually worth it if what comes out the other end has none of the texture that made customers recognize the business in the first place.
What taste actually protects.
Your brand owns small recognizable details, a color, a shape, a sound, a way of talking, and marketing research has a name for them: Distinctive Brand Assets. The framework comes from the Ehrenberg-Bass Institute, built on researcher Byron Sharp's work on how brands grow, with testing methods developed by fellow researcher Jenni Romaniuk. It scores these assets on two measures: fame, meaning how many people link the cue to your brand without being told, and uniqueness, meaning whether that link belongs to you alone or gets shared with competitors.
The framework's test is almost embarrassingly simple: show someone an unbranded color block, a silhouette of your packaging, or a few seconds of a sound, and ask which brand it belongs to. If they guess right, and they guess right more than they guess your competitor, you own something. That ownership is built slowly, one consistent ad at a time, and it is exactly what gets sanded off when a tool defaults to whatever looks clean and generic.
This is where taste earns its keep. Taste, in this context, isn't a feeling about what looks nice. It's the owner's accumulated knowledge of which specific details are theirs: the exact shade of green on the awning, the way the copy always undersells itself a little, the photo style that's slightly rough instead of polished. An AI tool has no access to that history. It will smooth toward the average unless someone tells it not to. Someone still has to catch the drift back toward generic before the ad goes out, which is why Knowing looks at an ad before a dollar is spent on it.
The cost of short-term thinking predates the tools.
It would be convenient to blame AI for a slide toward sameness and speed, but the slide started well before anyone typed a prompt into an image generator. Les Binet and Peter Field, two marketing researchers, have spent over two decades analyzing entries to the IPA (the UK's Institute of Practitioners in Advertising) Effectiveness Awards, and the trend line is not encouraging. As Marketing Week reported on their analysis, the average number of very large business effects reported by creatively-awarded campaigns fell to 1.4 in 2018, the lowest point in the award's 24-year run, down from a high of 1.9 a decade earlier.
The same reporting found that short-term campaigns, meaning campaigns designed to show results in six months or less, rose from about 10 percent of creatively-awarded entries in 2002 to 25 percent by 2018. Among all award winners the figure climbed to nearly 40 percent. Binet and Field's broader research points to a roughly 60/40 split favoring long-term brand building over short-term sales activity as the effectiveness sweet spot, and their data suggests most businesses invest well below that. That sweet spot is an average across a healthy marketing budget, not a rule for any single ad, and a business that needs sales this month can still make space for brand-building work over the year. AI tools arrived into a landscape that was already optimizing for this quarter over the next three years.
A fast ad that looks like everyone else's fast ad isn't actually faster. It just fails later instead of sooner.
That reframes the real risk. It isn't that AI makes bad ads. It's that AI makes it easier to keep making the short-term choice the industry was already making, at a pace that outruns the owner's ability to notice they're doing it.
What keeping your taste actually looks like.
Keeping taste while moving fast isn't about rejecting the tools or slowing back down to the two-hour version. It's about deciding, ahead of time, what in your ads is non-negotiable, and checking every fast draft against that short list before it goes live.
A plumber's business runs on a specific shade of orange on every truck and every flyer for fifteen years. An AI tool, asked to generate a Facebook ad, defaults to a generic blue because blue tests well in stock templates. Nothing about the ad is wrong on its own. It is simply not the plumber's orange, and a returning customer who has seen that orange for a decade will register, even faintly, that something is off.
That kind of miss is invisible to a tool working from a prompt. It is not invisible to the owner, if they're the one checking the draft against the two or three things their brand actually owns: a color, a phrase, a way of framing an offer. The check takes minutes. Skipping it means your ad ends up looking like your competitor's.
The discipline, then, isn't about being anti-AI or pro-AI. It's about treating speed as a tool for getting more drafts in front of your own judgment, not a reason to stop exercising that judgment. The owner who keeps a short list of what's theirs and checks every draft against it before publishing gets the speed without losing the recognition that speed was supposed to protect.
Frequently asked
Common questions.
Does using AI to make ads mean losing my brand identity?
Not automatically. AI tools default toward generic choices unless directed otherwise, so the risk is real, but it's manageable. Keep a short list of the specific colors, phrases, or visual habits that are yours, and check every AI-assisted draft against that list before it goes out.
What are distinctive brand assets?
They're the non-verbal cues, like a color, shape, sound, or visual style, that let customers recognize your brand without seeing your name. The Ehrenberg-Bass Institute's framework tests them on fame (do people link the cue to you) and uniqueness (is that link exclusive to you, not shared with competitors).
Is speed in ad production actually a problem?
Speed itself isn't the problem. Research from Field and the IPA shows ad effectiveness has been declining for two decades as short-term thinking rose, well before AI tools existed. The risk is that faster tools make it easier to default to short-term, generic choices without noticing.