AI adoption is rising — but not evenly
AI in marketing has moved past the novelty phase, but that doesn’t mean it’s been welcomed equally everywhere. A first-quarter 2026 survey of more than 100 marketing professionals found a pretty lopsided pattern: some channels are already using AI as part of normal work, while others still treat it like a tool they’ll get around to testing later. The split is especially visible in influencer marketing and connected TV, where adoption trails far behind the places marketers have already gotten comfortable experimenting.
Earlier research painted the same picture. Social media and retail media have seen much stronger AI use than influencer marketing and CTV, which makes a certain kind of sense once you think about the day-to-day work in each channel. Social and retail media tend to generate tidy signals. There’s plenty of performance data, lots of repeatable formats, and fast feedback. AI has room to sort, predict, and optimize there without stepping on too many toes.
Influencer marketing is a different animal. Roughly one in four marketers said they use AI for influencer work, which is hardly a stampede. CTV lagged even more sharply. Close to four-fifths of marketers said they aren’t using AI there at all. That doesn’t mean they dislike AI. It usually means the channel doesn’t feel like a natural fit yet, or at least not one they trust enough to hand over real decisions.
AI gets adopted fastest when the work is repetitive, measurable, and easy to correct. The messier the channel, the slower the rollout.
That gap matters because it shifts the conversation away from generic enthusiasm for AI and toward channel fit. People are no longer asking, “Should we use AI?” in the abstract. They’re asking, “Where does it actually help, and where does it start to get awkward?” In influencer marketing AI, the awkward part often sits right next to the thing that makes the channel valuable in the first place: a human voice that feels believable. In CTV, the hesitation comes from a different place. Streaming ads still carry the weight of a premium screen, bigger budgets, and creative that can look a little off if the machine gets too clever too quickly.
So the pattern isn’t really about whether marketers like AI or fear it. They already like it, at least in the places where it saves time and cleans up routine work. The real question is why some channels invite AI in with open arms while others keep it waiting at the door. That’s where influencer marketing and CTV get interesting, because the answer has less to do with hype and more to do with what each channel can tolerate, measure, and trust.
Why influencer marketers still worry about authenticity
Once you move past the general excitement around AI, influencer marketing gets awkward fast. The whole channel runs on a simple expectation: people want the creator to feel real. That doesn’t mean every post has to be unpolished or every caption has to sound like it was typed at 1:12 a.m. On a phone in the dark. It does mean audiences notice when a brand’s voice starts talking through a creator in a way that feels scripted, sterile, or a little too perfect.
In influencer marketing, trust is the product. If AI makes the message feel manufactured, the whole arrangement starts to wobble.
That helps explain why so many brands move carefully here. An April 2025 industry study found that nearly all brands steering clear of virtual influencers pointed to consumer trust concerns. That hesitation makes sense. A virtual creator can be visually polished, but polish is not the same thing as credibility. If a campaign depends on the audience believing that a creator actually uses the product, has an opinion, and has some degree of personal stake in the recommendation, then the bar is higher than it is for many other AI for ad campaigns. A brand can experiment with AI in media planning or reporting and still keep a straight face. Influencer content asks for something messier.
That doesn’t mean marketers are keeping AI out of influencer work entirely. They’re just drawing a line around where it enters. In practice, about three-quarters of the marketers who use AI in this channel rely on it for data analysis. That’s the safest place for it, frankly. If AI can sort campaign results, flag which creators moved product, or compare audience response across posts, it can save hours without putting words into someone else’s mouth. A little over half use it for content creation, and just over half use it for outreach. Even there, the more cautious teams seem to treat AI as a drafting or sorting tool rather than a ghostwriter with a ring light.
Beekman 1802 offers a good example of that middle path. The brand used AI alongside first-party CRM and Shopify data to identify customer subtypes that were not obvious from a glance at the sales sheet. Those groups now inform messaging and campaign planning. That is a very different use of AI than asking a machine to mimic a creator’s personality. It’s closer to pattern recognition. The brand looks at who buys, how they buy, and what seems to separate one set of customers from another, then uses that information to shape the brief before a creator ever gets involved.
Later, AI also helps Beekman 1802 match campaign briefs with human creators and predict how content might perform based on past engagement data. Again, the machine sits behind the curtain. It can narrow the field, compare past results, and suggest likely fits, but the actual influencer relationship still depends on judgment. Who sounds believable for this product? Who has the right audience? Who can say yes in a way that doesn’t sound like they were handed a script and told to smile?
That last question matters more as the creator economy gets crowded. Brands are working with more creators than they used to, especially smaller and mid-sized ones, and that makes manual campaign management harder to sustain. Sorting through dozens, sometimes hundreds, of potential partners by hand is slow, and it’s easy to miss people who would have been a strong fit. AI can help with the sorting, the scoring, and the follow-up. What it can’t do, at least not convincingly, is fake the human relationship that makes influencer marketing work in the first place.
That tension is probably why the channel remains cautious even as other areas move faster. Social posts can be optimized, priced, and measured with plenty of AI help. Influencer work asks for something more delicate: speed on the backend, restraint on the front end.
Creators are using AI to find work and move faster
If brands are nervous about AI making influencer marketing feel fake, creators have taken a much more practical view. For a lot of them, AI is just another tool in the stack, like scheduling software, thumbnail editors, or the half-dozen apps that already keep a creator’s day from collapsing into chaos. A Wondercraft report put the number at around eight in ten creators using AI at some point in their workflow, which is a pretty strong sign that the creator side of the internet has already made peace with the machine.
That doesn’t mean creators are handing over their voice to a bot and calling it a day. The better use cases are less glamorous and far more useful. AI can sort, rank, draft, summarize, and surface things that would otherwise get buried. In creator economy AI, that often means saving time on repetitive tasks so a person can spend more time doing the part only they can do: sounding like themselves. Nobody follows a creator for their inbox management skills. Well, almost nobody.
The smartest use of AI for creators is usually the least flashy one: find the opportunity, sort the noise, then get out of the way.
POP.STORE’s AI ECHO ME program fits that pattern neatly. It’s built to surface revenue opportunities, generate content, and help with fan engagement, which is basically a polite way of saying it tries to keep creators from missing money while they’re busy making the rest of the internet scroll. The platform can connect to a creator’s social accounts and sort through direct messages. Email support is planned for later, which makes sense, because once a creator has enough inbound messages, their inbox starts to look less like communication and more like a small disaster with notifications.
The real job here is triage. Not every DM deserves a reply, and not every message that sounds friendly is actually useful. AI ECHO ME is meant to sort out whether a note came from a brand or another creator, how large the sender is, and whether there’s a clear monetary offer attached. That last piece matters more than it might sound. Creators get a steady stream of messages that begin with enthusiasm and end with vague promises, “great exposure,” or some variation of “let’s circle back.” An automated system can at least separate the actual offers from the digital small talk.
There’s also a more human reason this kind of tool is finding an audience. POP.STORE’s CEO has said creators are overloaded, and that rings true if you’ve ever watched how their work actually happens. They’re making content, tracking comments, watching analytics, responding to partners, planning the next post, and trying to guess what the algorithm will do next. That last part is enough to make anyone suspicious of sleep. When all of that stacks up, a real opportunity can slip by simply because it arrived in the middle of a busy hour, or a busy week, or a busy life.
That’s where creator-side AI feels different from the anxiety around ad tech in brand-side campaigns. For creators, the technology is often less about replacement and more about relief. It can draft a reply, sort a message, spot a pattern, or flag a brand deal that might otherwise get lost under ten unrelated pings and a coupon code from someone’s cousin. Used well, it doesn’t flatten a creator’s personality. It just clears a little space around it.
And that’s probably the most realistic version of this whole shift. The strongest tools won’t turn creators into generic content factories, because audiences can smell that from a mile away. What they can do is help creators move faster, answer smarter, and catch the offers worth reading in the first place. In a business where timing matters almost as much as talent, that’s no small thing.
CTV is the toughest place for AI to break through
Among the channels marketers use every day, CTV has turned out to be one of the least welcoming spots for AI. In the survey, almost four in five respondents said they weren’t using AI in CTV campaigns at all. That looks very different from social media and retail media, where AI has already found a much firmer footing.
The gap makes sense once you think about how these channels were built. Social platforms grew up around data. They were designed to collect signals, test content, and adjust quickly. CTV came from a broadcast mindset first. It inherited the old TV habit of buying reach, packaging inventory, and working with broader audience segments, then added streaming layers on top. AI can still fit into that world, but it has to work a little harder to prove itself.
As CTV becomes more addressable, the setup changes. Marketers can use first-party data, attention signals, and contextual intelligence instead of relying only on broad programmatic guesses. That opens the door to AI in a few specific places: audience targeting, contextual targeting, creative generation, and media buying. In other words, AI is not just helping people buy more TV ads. It’s helping them decide which viewers to reach, what kind of content they’re watching, what creative asset to show them, and how to spend the budget without wasting half of it on guesswork.
CTV rewards disciplined automation more than flashy automation.
That distinction matters because streaming ads don’t give marketers the same freedom they get in open-web display or some social environments. A badly matched ad on TV tends to feel more expensive, more noticeable, and harder to ignore. So when AI enters the room, the bar is higher. The output has to look and sound like something a serious brand would actually air, not a rushed template with a logo slapped on top.
Amazon has started pushing into that space with tools that cover more than one step of the process. Its Complete TV product recommends ways to spread spend across Prime Video and other premium streaming inventory. That matters for advertisers trying to keep CTV buys from becoming a pile of disconnected placements. Amazon also offers AI Creative Studio and Audio Generator, which let brands produce video, image, and audio assets without starting every campaign from a blank page. For teams already stretched thin, that kind of marketing automation can shave off a lot of grunt work.
Roku is taking a similar path for smaller advertisers. Its Ads Manager uses AI partnerships to make TV advertising feel less like a locked club and more like something a direct-to-consumer brand or local business can actually try without building a giant media team first. That’s a big deal for smaller budgets, where one wrong decision can eat a painful chunk of spend. If AI can help with setup, targeting, and buying discipline, then CTV becomes less intimidating and a lot more practical.
Of course, there’s still a real worry about low-quality AI creative. Streaming ads live on a large screen, and bad copywriting or awkward visuals are much harder to shrug off there than in a scrolling feed. Still, CTV platforms usually have more control than social networks do. They can set tighter standards around where ads appear, how assets are delivered, and what formats are allowed. That doesn’t remove the risk, but it does give brands more room to keep things polished.
So CTV may be late to AI adoption, but it doesn’t look resistant in principle. It looks selective. Marketers seem willing to use AI where it helps them target better, buy smarter, and produce assets faster, while keeping a close eye on quality. That caution may slow adoption, yet it also fits the channel. Nobody wants their premium streaming ad to look like it was assembled in a hurry by a tired intern with one eye on lunch.
Where AI in advertising goes from here
A lot of the hesitation around AI in CTV has less to do with the model itself and more to do with the plumbing underneath it. Identity resolution is still fragmented. One vendor may know a household saw an ad, another may track a visit, and a third may claim credit for the sale. Put those pieces together, and the picture often looks fuzzier than marketers would like. Measurement lives in silos, too, which makes it hard to tell whether AI improved anything or just made the reporting deck look more polished.
The advertising tools that win won’t be the ones that sound smartest. They’ll be the ones that save time without making the final result feel risky.
That’s why the near-term use of AI is likely to stay mostly backstage. Targeting can get better. Editing can get faster. Campaign optimization can be less manual. Production teams can use AI content creation to draft variants, resize assets, or test copy before a human spends hours polishing the final cut. The work is useful, but it doesn’t ask brands to hand over the whole store.
A research project published in April 2026, with several Netflix employees involved, pointed in that direction. It introduced VOID, an AI tool that can edit video by removing objects and understanding how things relate inside a scene. That kind of capability matters because it deals with real production chores, not just novelty. If a tool can clean up footage, support versioning, or help a team move from one cut to another without starting over, it can save actual labor. That’s the sort of improvement marketers can measure without much hand-waving.
Still, TV creative is a finicky business. The biggest screen in the house tends to expose weak spots fast. A slightly off facial edit, a strange object removal, or a frame that looks vaguely wrong can distract viewers in seconds. The same caution that slowed interest in virtual influencers applies here in a different form: when the audience can sense something artificial, trust drops. For brands spending real money on premium inventory, that risk may outweigh the convenience of faster generation unless a person checks the work.
So the practical path forward seems clear enough. AI will keep spreading in advertising, but not as a magic replacement for strategy or judgment. It will sit inside targeting systems, editing tools, optimization platforms, and production workflows, where it can do boring work faster than a team of humans staring at a timeline at 9 p.m. If those tools reduce hours, trim waste, and make campaigns easier to manage, the payoff can eventually show up where advertisers care most: efficiency, cleaner execution, and better performance.
