The influencer selling to you might not be a person

A while back, a friend of mine who works in branding cornered me for a venting session.

He said influencers were getting harder and harder to find. Budgets were shrinking every year, quotes climbing every year. He'd send out a brief and hear nothing back; when he did lock down a slot, the results were a roll of the dice. Finally he blurted out: "At this rate, I'm going to have to split myself into three people."

I didn't rush to console him. I just handed him a report.

It was the latest benchmark report from Influencer Marketing Hub, which had asked 500 people who do influencer marketing for a living three questions: Do you use AI? Are you afraid AI will take your job? And which parts of the work would you hand over to AI?

Once the responses came in, everything boiled down to a single line:

You think you're competing with influencers. In reality, you're already competing with AI.

Don't rush to argue. The report has two big parts: how AI has wormed its way into the backend of influencer marketing, and how AI is now outright fabricating "people" to sell things. Let's take them one at a time.

1. Hand the backend grunt work to AI

What counts as backend work?

Writing briefs, drafting copy, vetting creators, organizing creator assets, sending templated messages, pulling data for reviews. The kind of work that's tedious to do and impossible to skip.

And these are exactly the tasks AI is best at. Look at the numbers: 48.7% of people doing influencer marketing already use AI "every single time." In 2025, a survey of 50-plus influencer-marketing operators found over 80% said AI had become part of daily life, not some new experiment.

Go one layer deeper into the tech stack and it gets more interesting. Asked which AI technologies they currently use, the top answer was natural language processing, at 49% — in plain English, getting the computer to write for you. After that came machine learning at 28.7%, deepfake technology for photorealistic short video at 24.3%, predictive analytics at 22.3%, and audience segmentation at 18.7%. Lay those numbers side by side and you realize that so-called AI influencer marketing stopped being some sci-fi project a long time ago; it's an assembly line that runs every day.

Take a tool as a case in point. Upfluence, a leading influencer-marketing platform, has plugged ChatGPT into its system for finding and reaching out to creators — even the opening lines are customized per person. These days, even the slog of "winnowing 500,000 influencers down to ten solid ones" is something AI does faster than a human.

But what truly unites marketers is a dividing line, and it's startlingly clear: vetting influencers' credentials, maintaining relationships, giving creative feedback, and negotiating contracts all stay in human hands.

AI is a backend accelerator, not a front-line decision-maker. What it takes over is the time-consuming work.

2. Where the money flows, the funding sheets will tell you

If you want to judge whether an industry has a future, don't look at the slogans first. Follow the money.

On the content-tools track, Jasper — which specializes in copy generation — raised the most. Synthesia makes avatars and video, Anyword does copy optimization, and each one raised more than the last.

Over on the "fabricating virtual influencers" track, the bets got even bigger:

• Brud, the company that made virtual influencer Lil Miquela famous, has raised $131 million in total;
• Xmov, based in Shanghai, wants to build an "infrastructure for the virtual world" — it has raised $147 million across four rounds, backed by Sequoia China and SoftBank;
• Superplastic has raised $20 million, led by Amazon.

$131 million, $147 million, $20 million. Put those numbers together and there's only one way to read them:

Capital is voting for AI influencer marketing — and betting heavily.

3. Scared, but using it anyway

The money talks, but people are still conflicted.

35.6% of marketers admit they're afraid AI will take their jobs. Yet a similar share of another group swears they're not worried at all — they've plugged AI straight into their workflow as a mainstay.

Now look at how the money is split. Four in ten marketing teams spend less than a tenth of their budget on AI; but two in ten teams have already committed four-tenths of their budget to AI. Two ways of operating in the same industry. And here's a number every brand should take note of: 43.8% expressed strong concern about the ethical issues lurking behind AI influencer marketing.

71.2% of respondents admit that AI really is better than people at "certain specific tasks." But ask what must be held back, and the answers are remarkably consistent: 42.2% say top-level strategic decisions still need humans, and 22.6% say creativity and emotional resonance are off-limits for AI.

See, what people fear isn't "AI replacing humans." It's "not having figured out yet exactly which part of me AI will replace."

Here's the math I gave my friend: before, one person had to keep the books and carry the weight. AI walks in, takes over the bookkeeping entirely, and frees you up to do the weight-carrying work with more force and precision. The work didn't shrink — the structure changed.

AI takes the repetitive work; people hold on to the work of judgment and emotion.

Whoever figures out that dividing line first has nothing to fear.

4. Have the people who actually use AI reaped the rewards?

By now you're probably wondering: the ones who actually use it — does it work?

Six in ten have tasted the benefits. 36.3% say things are "noticeably better," and 23.5% say "slightly better." On the flip side, 8% say noticeably worse, 12.7% say slightly worse, and 19.5% just shrug — no change, stuck in place.

The same AI produces wildly different results. What's the difference? How you use it. Plugging it into the right scenario, wiring it into the workflow, keeping expectations level — that's one approach. Treating it like a wishing machine is another.

AI isn't magic; it's a real, hard-nosed overhaul of your processes.

When marketers pick a tool, what matters most? First is "letting AI generate content on its own" — 46.6% rank it first. Next comes data analytics (33.5%) and audience targeting (28.3%). As for "whether I can train it to follow my lead," that ranks fourth, at 22.7%. Content is always the No. 1 productivity engine.

When they overhaul their setup, marketers want one and the same thing: see the data first, know the outcome before it happens. Asked what they most want AI tools to improve, 41.3% asked for better predictive analytics — before spending every last dollar of budget, they want to know roughly how much of it will come back.

The tighter the budget, the bigger the win for whoever can "do the math" first.

One more number: 77.4% of marketers have fought with their AI tools and hit technical glitches. But after all the arguing, they're still using them.

Cursing it one minute, unable to live without it the next — now that's a real tool.

5. AI goes straight to work as an influencer

Tools aside, what comes next is the thing that reshuffles the entire industry.

AI isn't just an assistant anymore. AI is now outright playing the "person" — the virtual influencer.

What's a virtual influencer? I'll paint the picture in one sentence: a person conjured up by AI, with a name, a face, a personality, and hundreds of thousands to millions of followers, who can post, engage, and sell.

First, adoption. In the report, 59.6% of marketers already use virtual influencers or plan to; 24.7% haven't touched them, and 15.5% are mulling it over. Among those who've used them, more than half report a "very positive" experience.

Ask whether "being able to control it" matters, and most people's answer is: very much. 42.6% called it "very important," and 33.5% went straight to "extremely important." Why do they care so much? Because the biggest advantage of a virtual influencer is that you have total control over what it says. 31.7% ranked that as the top benefit; second, at 29.1%, was "never off the clock, 24/7." Real celebrities throw tantrums, say the wrong thing, and get jet-lagged. A virtual influencer doesn't.

And the top names? The hottest virtual influencer is Lu do Magalu, from Brazilian retailer Magalu, with upwards of six million followers — first by a long shot. The fastest-growing one has a name that sounds like a joke: Nobody Sausage — a virtual sausage with an absurd look. Think you'd look down on it? Hugo Boss used it in its Spring/Summer 2022 campaign. Netflix even had it re-enact the trailer for Red Notice — engagement jumped 62%, and average view-through time hit 1.53x the original.

A sausage can carry a global marketing push. Read all that and it's hard not to laugh.

One more number: the average engagement rate for virtual influencers is 2.84%, while real humans manage only 1.72%. Yes, you read that right — AI influencers get engagement rates roughly 60% higher than real people.

Who's watching? 65.5% of their followers are women, and teens aged 13–17 make up 11.6%. The content core is Instagram Reels — short, fast, punchy.

Geographically, the virtual influencer home base is the US, with plenty in Japan, the UK, Brazil, and South Korea too. It never gets jet-lagged — every time zone is working hours. That's the kind of freedom a real influencer whose calendar is booked through next March could never dream of.

AI influencers aren't competing with humans over "who's real." It's about who's fresher, faster, and savvier.

6. Will consumers open their hearts to "fake people"?

You might say: it's fun to watch, sure, but when it's time to actually open their wallets, people will only pay real money to real humans, right?

In 2022, the international journal European Journal of Marketing published a serious study on exactly this, titled "Distorted Influence." The conclusion is rather clever:

Consumers do see AI influencers as less trustworthy and psychologically further away. But on the two questions of "would I follow" and "would I share," AI influencers and real people end up in a dead heat.

Why? Because what people care about has never been "is it real" — it's "is it entertaining."

Consumers may not trust the influencer, but they're happy to watch it.

It's entertaining, all right. But whether it can actually sell is another matter. Which brings us to the next question: what kind of product can a virtual influencer actually move?

7. There's no "best" influencer, only the best-matched one

In January 2023, an experiment was run: a real influencer and a virtual influencer each promoted the same lineup of products — Lancôme makeup, a Samsung Bluetooth speaker, and Calvin Klein face cream. The participants were a group of German women.

The results were fascinating, with a double twist:

• On "likeability," the human wins: 22% found the human-endorsed ad more endearing, while the virtual influencer only scored an 11.2% "more novel."
• But when it came to "would you actually buy," the story flipped: for the speaker, people preferred the virtual influencer to sell it; for the face cream, the human influencer was more persuasive.

Same product, different face, and the buyers' sentiments swing completely.

There's no strongest influencer — only the one that fits best.

8. Which lever actually moves the purchase?

Here's the make-or-break question: when a virtual influencer goes to work, what actually wins people over?

Most people guess "good looks." Think again.

Here's what the data says: when consumers watch a virtual influencer hold up a product, what matters most is "realism" — 38.6% rank it first. Second is "whether the product and the influencer are a good match," at 25.1%. Only 8.8% genuinely care about "looks."

What consumers want isn't a perfect person; it's someone who looks real on the surface.

Here's something even more striking: 52.4% of industry professionals believe the versatility of AI influencers will rewrite the script for the future of marketing and entertainment. Think about it — it doesn't sleep, doesn't need scheduling, never misspeaks, and never implodes over a scandal. What flesh-and-blood human anywhere in the world can compete with that?

As for the remaining four in ten, they're still on the sidelines. They haven't tried it firsthand, and they haven't been burned either. Their next move is your next opportunity.

9. Wrapping up

The story has one last line, and it brings us back to that friend from the start.

I asked him: you spend every day trying to figure out how to split yourself into three people — you've got it backwards. What you should be thinking is how to walk through every task again: which jobs would AI do better than you? And which jobs, the moment you hand them over, would mean you can never really be yourself again?

He paused for a few seconds. "Then I'll just hand half of it to AI for now."

See, that's what it looks like when someone finally does the math. What's happening isn't "AI replacing people." It's "people who use AI replacing people who don't."

If you're worried about AI taking your work, the people you should really watch out for are the peers who've learned to make AI grab time right alongside them.

I don't know what this industry will look like three years from now. But I do know this: the people who do the math right will still be at the table.

When that day comes, I hope you're the one sitting across the table, playing cards alongside AI. And I hope you've already got that half of your hand firmly in your grasp.