AI Ad Operator
For two decades, boutique owners running paid ads have had three options: hire an agency, hire in-house, or run ads themselves. Each option has the same problem at its core — running profitable ads for a boutique is daily work. Checking which products are converting this morning, deciding whether yesterday's ad set deserves more budget today, noticing that a hero product is about to sell out, creating a new ad to replace the one that is burning out by Thursday. That work does not get smaller as a boutique grows. It gets harder.
The agency model assumed humans had to do that work, so agencies priced it like human work: $1,500 to $3,000 per month plus a percentage of ad spend. For boutiques spending $100,000+ per month on ads, that math works. For boutiques spending $1,000 to $50,000 per month (most of them), the math has never worked — which is why most boutique-agency relationships end in frustration within 12 months.
The DIY option assumed boutique owners had time to learn Meta Ads Manager and execute daily. Most owners who try spend 5 to 10 hours per week on it once past the learning curve. That is a part-time job, on top of running the actual store.
A new category replaces both. The AI ad operator. Lenny is software that performs the daily work of a media buyer for boutique brands. It does six things every day: monitors every active ad against performance baselines, catches ads that are slipping before they cost real money, surfaces ads working better than expected, connects ad performance to inventory data so it knows when a winner is about to sell out, generates the next round of creative based on what is currently working, and presents all of this as one-click actions.
The boutique owner stays in control of every decision. The AI does the watching, the analysis, the creative drafting, and the recommendations. The owner approves, edits, or rejects. The strategic work — brand voice, creative judgment, seasonal planning — stays with the owner. It was always the human's job. The clicking and watching was not.