AI Influencers: What Brands Are Buying

AI influencers are computer-generated characters that brands pay to promote products across social media, typically through sponsored posts, brand partnerships, and content collaborations that look, on the surface, almost identical to those produced by human creators. They have followers, they have aesthetic identities, and some of them generate millions of engagements per month without ever drawing breath. Whether that represents a genuine shift in how influence works, or a very expensive way to avoid dealing with real people, depends entirely on what you think influence is actually for.

I have spent more than twenty years watching new formats arrive in marketing with the same promise: more reach, less friction, better ROI. Paid search delivered on that promise. Social media advertising largely did. NFTs did not. The honest answer on AI influencers is that we are somewhere in the middle, and most brands evaluating them are not asking the right questions first.

What Is an AI Influencer, Exactly?

An AI influencer is a fictional persona rendered using computer graphics, generative AI, or a combination of both, designed to exist and operate on social media platforms as though it were a real person or character. The most prominent examples, such as Lil Miquela, were built before generative AI made the process dramatically cheaper. They had backstories, opinions, musical releases, and brand deals with companies including Calvin Klein and Prada. They also had the kind of carefully managed ambiguity about their nature that now looks like a strategic choice rather than an oversight.

More recent AI influencers have been built faster and at lower cost, using generative tools to produce images, video, and even voiceovers without a human model involved at any stage. Some are branded explicitly as AI. Others are not. That distinction matters more than most brand teams currently appreciate.

The mechanics vary. Some AI influencers are operated by agencies or studios with dedicated creative teams managing content calendars, brand partnerships, and community engagement. Others are closer to automated content pipelines with a visual persona attached. The quality gap between those two approaches is significant, and it is not always visible from the outside.

If you want to understand where AI influencers fit within the broader shift in how generative tools are reshaping marketing, the AI Marketing hub at The Marketing Juice covers the landscape in detail, including where the technology is genuinely useful and where it is being oversold.

Why Are Brands Interested in Them?

The pitch for AI influencers tends to cluster around three arguments: control, consistency, and cost. None of them are as clean as they sound.

Control is the most frequently cited reason. Human influencers are unpredictable. They express opinions, make mistakes, get into controversies, and sometimes turn out to hold views that are incompatible with a brand’s positioning. The appeal of a synthetic persona is that it does what you tell it to do, says what you need it to say, and does not post anything unexpected at two in the morning. That is a genuine advantage, but it is also a description of a brand channel with a face on it, which is not the same thing as an influencer relationship.

Consistency is related. AI influencers do not age, do not change their aesthetic, and do not renegotiate their rates when their following grows. For brands that have invested in a visual identity built around a persona, that stability has real value. Fashion and beauty brands have been particularly interested in this because their visual language is so tightly managed already.

Cost is where the pitch starts to break down. The upfront investment in creating a credible AI influencer is not trivial. Building a character with enough depth and visual consistency to attract a genuine following requires creative resources that most brands are not accounting for in their initial estimates. The ongoing content production, community management, and brand partnership work still requires human oversight. The savings are real in some areas and largely illusory in others.

I have sat across the table from enough agency new business pitches to recognise when a cost argument is being used to sell something that the buyer actually wants for a different reason. The brands that are genuinely excited about AI influencers are not primarily motivated by cost. They are motivated by the idea of owning the relationship entirely, and that is a different conversation.

Do AI Influencers Actually Drive Engagement?

The engagement numbers on some AI influencer accounts are legitimate. Lil Miquela has accumulated millions of followers across platforms and generated engagement rates that many human influencers would be satisfied with. More recent AI personas have also attracted genuine followings, particularly in gaming, fashion, and entertainment-adjacent categories where audiences are already comfortable with fictional characters as cultural figures.

But engagement is not influence, and conflating the two is one of the most persistent errors in influencer marketing regardless of whether the creator is human or synthetic. People follow AI influencers for a range of reasons: novelty, aesthetic appeal, entertainment, curiosity about the technology itself. Those motivations produce engagement behaviour, including likes, comments, and shares, that looks similar to the engagement generated by trusted human creators but functions differently in terms of purchase intent and brand association.

The question worth asking is not whether people engage with AI influencer content. It is whether that engagement translates into the commercial outcomes the brand is trying to achieve. When I was managing large-scale paid media accounts, the discipline we applied was simple: follow the money back to the outcome, not to the proxy metric. Engagement rates are a proxy. Conversion, brand lift, and incremental revenue are the outcomes. The gap between those two things is where a lot of influencer marketing budget disappears, and AI influencers are not immune to that problem.

There is also a category-dependency here that tends to get glossed over. AI influencers have shown the most commercial traction in categories where aspiration and aesthetics do most of the work: luxury fashion, streetwear, beauty, gaming peripherals, and entertainment. In categories where personal recommendation and authentic experience matter more, such as health, finance, parenting, and food, the synthetic nature of the persona is a structural disadvantage that engagement numbers cannot compensate for.

The Disclosure Problem Is Not Going Away

The regulatory environment around influencer marketing has been tightening for years, and AI influencers are now explicitly in scope. In the United States, the FTC’s guidelines on endorsements and testimonials require clear disclosure when a relationship exists between a brand and a promoter. The question of how those rules apply to a synthetic persona that is openly or ambiguously artificial is still being worked through, but the direction of travel is clear: regulators expect consumers to know when they are being marketed to, and they expect that disclosure to be unambiguous.

Several countries have already moved further. The EU’s Digital Services Act creates obligations around transparency in advertising that extend to AI-generated content. The UK’s Advertising Standards Authority has been increasingly active on influencer disclosure more broadly. Brands operating across multiple markets are handling a patchwork of requirements that is only going to get more complex.

The brands that are treating disclosure as a box-ticking exercise are making a mistake. The reputational risk of being seen to deceive consumers about the nature of an AI persona is substantially higher than the reputational risk of being transparent about it. Some of the most successful AI influencer campaigns have leaned into the synthetic nature of the persona as a creative asset rather than something to be minimised. That is the smarter position, both ethically and commercially.

Generative AI video tools are making this more complicated, not less. As HubSpot’s overview of generative AI video tools illustrates, the capability to produce photorealistic video content of synthetic personas is now accessible at a price point that puts it within reach of mid-sized brands, not just enterprise players. The production quality ceiling is rising fast, which means the disclosure conversation is becoming more urgent, not less.

What AI Influencers Cannot Replace

The most durable thing a human influencer offers is not their following. It is their credibility within a specific community, built over time through consistent, authentic engagement with a real audience that has chosen to trust them. That credibility is transferable to brand partnerships when the relationship is managed well. It is also fragile, which is why influencer marketing requires genuine relationship management rather than transactional contracting.

AI influencers do not have credibility in that sense. They have recognition, aesthetic identity, and in some cases a kind of parasocial affection from their audiences. But the mechanism by which a human influencer’s recommendation carries weight, which is the implicit endorsement of someone whose judgement you have come to trust through repeated exposure, does not operate in the same way when the audience knows or suspects the persona is synthetic.

This is not a permanent limitation. Cultural norms around what constitutes authentic endorsement are shifting, and younger audiences in particular have a more sophisticated and less binary relationship with the distinction between real and synthetic personas. But it is a current limitation that brands should be honest about rather than hoping the engagement metrics will paper over.

There is also the question of cultural intelligence. Human influencers are embedded in the communities they serve. They pick up on shifts in language, mood, and values in real time because they are living inside those communities. An AI persona managed by a creative team is always operating with a degree of latency. The team observes, interprets, and responds, but they are not the community. In fast-moving cultural contexts, that lag is a meaningful disadvantage.

When I was building out the influencer and content capability at an agency, the thing we kept coming back to was the difference between reach and resonance. Reach is measurable. Resonance is the thing that actually moves people. AI influencers can generate reach. The resonance question is more complicated, and it is the one that determines whether the investment makes commercial sense.

Where AI Influencers Might Actually Make Sense

There are genuine use cases where an AI influencer is a rational choice rather than a novelty play. They are narrower than the current enthusiasm suggests, but they are real.

Brand-owned personas are the clearest case. A brand that wants to create a consistent visual character to anchor its social content, without the complexity of managing a human talent relationship, can use an AI persona as a creative vehicle rather than a credibility vehicle. The persona becomes a brand asset rather than a third-party endorsement. That is a different brief, and it is one where the synthetic nature of the character is a feature rather than a limitation.

International market expansion is another case worth considering. Human influencers are culturally specific in ways that create real complexity when a brand is trying to maintain consistent positioning across multiple markets. An AI persona can be adapted for different cultural contexts without the contractual and logistical overhead of managing multiple talent relationships. That flexibility has genuine operational value for brands with complex international footprints.

Gaming and entertainment categories represent a third legitimate use case. Audiences in these categories have well-established relationships with fictional characters as cultural figures, and the line between synthetic and real is already blurred in ways that make an AI influencer feel contextually appropriate rather than jarring. The use of AI tools to build and manage content channels is already common in these categories, and AI influencer personas fit naturally into that ecosystem.

What these use cases have in common is that they are not trying to replicate what human influencers do. They are doing something different with a different mechanism. That clarity of purpose is what separates the brands using AI influencers intelligently from the ones chasing a trend.

The Production Reality Behind the Persona

One of the things that gets lost in coverage of AI influencers is how much human work goes into running them. The most successful AI influencer accounts have dedicated creative teams managing content strategy, visual production, brand partnerships, community engagement, and the ongoing development of the persona’s identity. The AI tools reduce certain production costs, but they do not eliminate the strategic and creative labour that makes an influencer account worth following.

Generative AI has made it dramatically cheaper to produce images and short-form video content of synthetic personas, and tools for automating parts of the content workflow are becoming more capable. Semrush’s analysis of AI content strategy gives a useful picture of where automation genuinely reduces workload and where it creates new quality control challenges. The same dynamic applies to AI influencer content production: the volume ceiling has risen, but the quality floor still requires human judgement to maintain.

There is also the question of what happens when something goes wrong. Human influencers have PR teams, agents, and personal judgement to draw on when a situation requires a response. An AI persona has a management team that is always one step removed from the audience. The response time and authenticity of that response are structurally different, and in a crisis, that difference matters.

Early in my career, I learned a version of this lesson in a different context. When I built my first website because the budget for a professional developer did not exist, I thought I was solving a production problem. What I was actually learning was that the production is inseparable from the strategy. You cannot outsource the thinking. The same is true of AI influencer management: the technology handles the rendering, but the judgement still has to come from somewhere.

How to Evaluate an AI Influencer Partnership

If you are a brand marketer evaluating an AI influencer partnership, the framework is not dramatically different from the one you would apply to any influencer investment. The questions are the same. The answers require more scrutiny.

Start with the audience. Who follows this persona, and why? What is the demographic and psychographic profile of the following, and how well does it map to your target customer? Request platform analytics rather than accepting screenshots. Understand the geographic distribution of the following, because AI influencer accounts sometimes have audience compositions that look good in aggregate but are geographically concentrated in markets that are irrelevant to your commercial objectives.

Then look at the engagement quality. Comments on AI influencer posts tend to be a useful signal. Genuine engagement looks different from bot-driven or curiosity-driven engagement, and the comment section is where that distinction is usually visible. Look for evidence that the audience is actually interested in the content rather than just reacting to the novelty of the persona.

Ask about disclosure practices and ensure they are legally compliant in your markets. This is not optional, and it is not something to leave to the AI influencer’s management team to handle without your explicit sign-off. Your brand is on the content. The compliance obligation is yours.

Finally, define your success metrics before you commit, not after. The temptation with influencer marketing, human or AI, is to measure whatever the platform makes easy to measure and call it success. Engagement rate is easy to measure. Incremental revenue, brand lift, and new customer acquisition are harder. Decide in advance which of those outcomes you are trying to achieve and build the measurement framework around that, not around the metrics the platform surfaces by default.

The broader question of how AI tools are reshaping marketing measurement and strategy is something I cover regularly across The Marketing Juice’s AI Marketing section, including the areas where the technology is genuinely changing what is possible and the areas where the hype is running well ahead of the evidence.

The Broader Influencer Marketing Problem AI Does Not Solve

It is worth stepping back from the AI-specific question for a moment, because a lot of the interest in AI influencers is driven by frustration with influencer marketing more broadly. Measurement is difficult. Attribution is contested. The relationship between engagement and commercial outcome is unclear. Talent management is complex and sometimes expensive. These are real problems, and they are driving brands to look for alternatives that feel more controllable.

But AI influencers do not solve the measurement problem. They do not solve the attribution problem. They make the talent management problem simpler in some ways and more complex in others. The underlying challenge in influencer marketing, which is demonstrating that the investment is generating commercial value rather than just social metrics, remains exactly as hard with a synthetic persona as with a human one.

The brands that are getting the most from influencer marketing, AI or human, are the ones that have done the harder work of connecting their influencer activity to business outcomes. They have built measurement frameworks that go beyond platform analytics. They have defined what success looks like in commercial terms. They have been honest about what influencer marketing can and cannot do within their specific category and customer base.

I have seen this play out repeatedly across different formats and channels over two decades. The format is rarely the problem. The discipline around how you use it almost always is. AI influencers are a format. The discipline question applies to them exactly as it applies to everything else.

The data on generative AI adoption in marketing shows that adoption is accelerating across almost every function, but the gap between adoption and effective use remains significant. AI influencers are a good example of that gap: the technology is available, the case studies are accumulating, but the strategic clarity about when and why to use them is still catching up.

What Comes Next

The technology behind AI influencers is improving faster than the commercial frameworks for using them. Photorealistic video generation is becoming accessible at scale. Voice synthesis is increasingly indistinguishable from human speech. The production quality ceiling for synthetic personas is rising every quarter, which means the visual and audio distinction between AI and human influencers is going to become progressively harder for audiences to detect without explicit disclosure.

That trajectory has implications for how brands, platforms, and regulators approach the space. Platforms are going to face increasing pressure to implement disclosure requirements at the infrastructure level rather than relying on creators and brands to self-disclose. Regulators are going to move from guidance to enforcement as the technology becomes more mainstream. Brands that have built their AI influencer strategies on ambiguity are going to find that position increasingly untenable.

The brands that will be best positioned as this space matures are the ones that have been transparent from the start, that have built genuine audience relationships through their AI personas rather than just accumulating followers, and that have connected their influencer investment to measurable commercial outcomes. That is not a technology question. It is a strategy question, and the answer looks the same whether the influencer is human or synthetic.

The tools available for building and managing AI-driven marketing workflows are evolving rapidly, and staying current on what is actually useful versus what is noise is a genuine challenge. Resources like Ahrefs’ AI tools webinar series and Moz’s coverage of AI tools for practitioners are useful for keeping a grounded view of where the technology is genuinely adding value.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions