Perplexity Says Its Computer Agent Ran $30K of Meta Ads for Get Maine Lobster and Beat the Brand's Agency
Computer managed $30k in Meta ad spend for Get Maine Lobster, outperforming their agency. Compared to their usual ads, Computer delivered 3.5x revenue with 63% higher ROAS. It automated creative analysis, ad production, and media buying. Read more:
Context
A paid ad campaign with real revenue numbers is a concrete test of whether AI agents can do marketing work that companies now pay agencies for, though the results come from Perplexity itself.
Computer is Perplexity's agent product that carries out multi step tasks on the web for users. Earlier in the Nov 3 deck, Perplexity promoted Computer for ecommerce work. Get Maine Lobster is an online seafood seller.
Deep dive
Perplexity (@perplexity_ai) posted on Nov 4 ET that its Computer agent managed $30,000 in Meta ad spend for Get Maine Lobster and outperformed the company's agency.
According to Perplexity, compared with the brand's usual ads, Computer delivered 3.5 times the revenue with 63% higher return on ad spend. Perplexity says the agent automated creative analysis, ad production and media buying. The post links to a case study page and includes a video.
These results come from Perplexity and its customer, not from an independent audit. The post does not say how long the campaign ran, what the agency's budget or targeting was, how the comparison period was chosen, or how much human review went into the ads.
The case study follows Perplexity's push to sell Computer for business tasks such as ecommerce, which it promoted the day before. Advertising is a large market where agencies are paid to plan, create and buy ads, so a claim that an agent beat an agency on revenue is a strong selling point.
The post passed 120,000 views by the time of this scan.
What to expect next
Watch for more case studies from Perplexity and for independent marketers testing Computer on their own ad accounts. Ad platforms such as Meta may set rules for agents that buy ads on behalf of businesses. Agencies will likely respond by adopting similar agents. Details on campaign length and human oversight would help judge how much of the result came from the agent.