The Influencer Doesn't Exist. The Money Is Real.

A few days ago I was scrolling social media and came across a blogger with a sizable following. She posted gym photos with a perfectly ordinary caption.

I tapped into her profile and kept scrolling. The further I went, the more something felt off.

This person does not exist.

She is a "virtual influencer." Drawn by a computer.

You might laugh: who watches a fake influencer? Plenty of people, actually. And not just a fringe few. One survey of U.S. consumers found that 58% of them have followed at least one virtual influencer, and 40% have bought something a virtual influencer recommended. By the estimates at the time, the market had already surged into the tens of billions of dollars.

Honestly? Wild. Fake people hawking products — and it's turned into a real business.

Today let's take this apart properly: Why do brands use them? Why does it work? And the most crucial question of all — trust.

What Exactly Is a Virtual Influencer?

It's a persona that looks like a real human being but entirely does not exist.

The early methods were crude: take a real person's photo as the base, paste a computer-generated face on top, and make sure it's the same face in every picture. Then large generative models arrived, images and video could be generated from scratch, and the crude method got upgraded into an assembly line.

Behind her might be a design studio, a media team, or the brand itself.

So why would a brand skip real humans and use a fake one?

Think about how hard it is to find a real influencer. What you want might be "someone who lives in Brazil, loves scuba diving, and also knows coffee." Geography, demographics, psychographics — everything has to line up. Does such a person exist? Possibly. Can you find them? Not necessarily.

A virtual influencer? Whatever configuration you want, just draw one up.

Even better: control. With a real influencer, you've paid the money, but how they talk and how much they say — you can't fully control it. And if one day they have a meltdown, the brand takes the blame alongside them. Virtual influencers don't melt down. They don't get DUIs, don't cheat on their partners, don't post unhinged rants in the middle of the night.

A real person comes with every risk of being human; a fake person comes with every obedience of not being one. That's the most seductive line on the brand's ledger.

Build One, or Partner With One?

Brands face two paths.

The first: build your own. It sounds lovely, but there's one fatal catch: nobody wants to follow an account that only posts ads. Think about it — a virtual human who talks about nothing but Coca-Cola, Nike, and Delta Air Lines all day. Would you follow her? No.

The sense of being a real person comes precisely from the "useless" content. Where she went on the weekend, what she ate with friends — daily life that has nothing to do with the brand. That's what makes her seem human.

The second: partner with an existing virtual influencer, paying for placements just like hiring a real one. Most brands today take this route.

But whichever path, they all end up answering the same question: was the money worth it?

Which brings us to the most interesting part of this research.

The Trust Switch Inside a Photo

A marketing professor who has studied word-of-mouth for nearly two decades had a team do something: scrape a large number of virtual influencer posts, run the images themselves through computer vision, and see which visual features drive engagement.

The finding is counterintuitive.

When "other people" appear in a virtual influencer's photos, engagement improves significantly. Especially when they look like real humans.

Posing with friends at a restaurant, having a training buddy next to her at the gym. That one detail alone lifts likes, comments, and shares.

Why?

Trust. A virtual influencer's built-in flaw: she can't actually have worn those shoes, can't actually have eaten at that restaurant. You know it. I know it. So her starting level of trust is far lower than a real influencer's.

So what does the person viewing the post do? Our brains hunt for clues. Clues of what? Evidence that "she seems human."

A real human friend appearing in her photo is evidence. Oh, she has social relationships, she has a life, she seems human. If she seems human, she can be trusted. If she's trusted, people engage.

And this "companion" must look like a real human. If standing beside her is another obviously cartoonish virtual character, the effect is discounted. Because consumers see through it instantly: this is still a fake person's world. A fake person can't vouch for another fake person.

A group photo is, in essence, a witness statement.

Consumers Aren't That Easy to Fool

At this point you might worry: so just draw a cute persona, photoshop a couple of "friends" in, and the money rolls in?

Not so fast.

Consumers are savvy. However adorable a virtual persona is, it can't dodge the fundamental question: has she actually brushed her teeth with the toothbrush she recommends?

This is exactly the trouble word-of-mouth runs into in the AI era. Twenty years ago, why was word-of-mouth credible? Because the recommender wasn't paid — his "it works great" was heartfelt. Then the influencer economy arrived, you knew he might have taken money, and trust took a discount. Now AI content has arrived, and even the "person" itself is fake.

For brands, the choice actually becomes clearer. If you want scale and control, accept the virtual influencer's low starting point of trust, then pull out every trick — say, one perfectly placed group photo — to stitch the "human touch" back in. If you want authenticity and word-of-mouth, find real everyday users and let them genuinely use the product and genuinely speak up. The latter are quieter, but every word of it lands hard.

Scale and trust sit on opposite ends of a seesaw. Every brand has to pick which end to sit on.

One Last Thing

That professor, after two decades studying word-of-mouth, says this is the most interesting period of change they've ever seen.

I'd argue the core of the change hasn't changed at all. From neighbors' gossip, to influencers' short videos, to computer-drawn influencers — the medium has cycled through generation after generation.

What people are really doing has always been the same thing: sifting through information for the answer to "should I trust you?"

A virtual influencer learning to bring "friends" into her photos is answering that question.

Who will answer it next? I don't know. But next time you scroll past an especially flawless influencer photo, take a second look at the person beside her.

She might be drawn, too.