Cameo introduces next-gen AI search to help you find the right celebrity for your message

Cameo—the hire-a-celebrity-to-talk-to-my-friend service—has a new way to search for the perfect person to record a clip. And, yes, it includes natural language search courtesy of generative AI.

Now users who find themselves lost in Cameo will automatically see a pop-up assistant offering help. If they want, they can enter into an AI conversation to help find the right celebrity for the job.

In recent years, Cameo has struggled to live up to its unicorn ambitions, but it is still very popular, with 50,000 celebrities and influencers of varying degrees on the platform. As Dom Scandinaro, CTO at Cameo, explains, with such a large number of people, it becomes difficult to find the right person you want to hire. That issue is exacerbated by the fact that 80% of Cameos are someone else’s gift—meaning that Cameo can’t simply profile your preferences and, in fact, its buyers may not have any information about the artist, athlete, or actor they’re looking for.

“[Discovery] it’s a big part of the field,” said Scandinaro. “We’re pretty bad about it today, I’d say.”

Scandinaro says that LLMs like ChatGPT seem easy for companies to use, in part because they don’t have the mindset to try. However, following his meditation on the technology over several months, he found it difficult to push the tool beyond a gimmick. “If you try to apply your company’s data to it and cover the LLM, that’s when it becomes more difficult,” he says.

That leaves room for companies like Loka, which is a next-gen AI fixer, to step in and help—as they did with Cameo. Over the past few years, Loka has launched hundreds of projects for companies with backgrounds ranging from entertainment to healthcare to energy. As a close (but not exclusive) partner of Amazon Web Services, Loka works as a technical and UX consultant, integrating the company’s data stored in the cloud and AI models that can make the most of it. With Loka, Cameo was able to integrate its data stored in the cloud with multiple LLM models for testing.

“I think it’s straightforward, you know, to point a model at a particular set of data, ask some questions about it and get the right answer. But I think if you’re really looking for something unique, personal to a particular situation, and not something that you ask every way the question, you can get the same result,” said Emily Kruger, Loka’s chief innovation officer. “That takes a lot of repetition.”

In theory, I understand everything Cameo and Loka said. LLMs can be surprisingly difficult to navigate—even Google search results are iffy. But in practice, what the team introduced to Cameo today doesn’t seem to improve the search bar.

I’ve tried a few chats—I’m pretending to be a dumb dad trying to find the right influencer for my kid on Twitch, or a clueless husband who wants to get a cookbook star to record a message for my spouse. What I found was that the conversation felt very race-focused—rather than really digging into someone’s interests or trying to find celebrities I might try to profile, it asked where my donors came from and quickly pushed the amount I wanted to have. use and how quickly I needed my recording. The results were generic, and no better or faster than what I could find directly from the search. I was very upset when I said I was looking for someone who works NYT Cooking or Bon Appetit. AI insisted that he had no one from him Bon Appetit. In fact, Cameo does. You had one job, Cameo AI!

This tool would be perfect if it was a chatbot coming out in 2018. It’s logically disappointing. I think the problem is, with services like ChatGPT and Perplexity setting such a high standard for conversation, discovery, and search—even if it’s wrong—anything short of that execution feels oddly outdated. It’s like how when you take a really good dark chocolate bar, a Hershey’s bar never tastes the same.

I do not point this out as trivial; I bring this up because it is proof that AI may not be the solution about your company’s biggest design problem, and that for AI to really move the needle on experience, we need to keep thinking beyond absolute information.

“I think for us, you know, we’re not in the health care or financial business, so we have a lot more flexibility to move things quickly, so we can learn to be fair and you know, 80% of the tests we have. we fail to find 20% that will make us successful immediately,” said Scandinaro.

The good news is that, indeed, this is a low-end app, and I have no doubt that Loka and Cameo will find this experience works better over time. But until then. . . there is always a good old traditional search.