- Is it really worth testing AI influencers in your company's campaigns?
- What do consumers really think about AI influencers?
- Do AI influencers build or destroy brand trust?
- Which brands have already used virtual influencers, and what can we learn?
- How much do AI influencers cost compared to human creators?
- Why everyone is talking about it now: are the Italian and European markets really accelerating?
- Three real-world scenarios: how does the choice change for different Italian companies?
- How do you launch an AI influencer test without burning audience trust?
- What mistakes are most common when testing AI influencers?
- How do I know if the AI influencer campaign is really working?
- So, are AI influencers worth it for your company?
- How we see it at SHM
- Frequently Asked Questions
- 45% of consumers wouldn't engage with an AI influencer and wouldn't trust its recommendations: the reputational risk is real and measurable.
- Trust works by category: fashion and gaming hold up better, while skincare and finance are a minefield because they require lived experience.
- Before launching, you need three things: a clear disclosure, sentiment KPIs (not just likes), and a benchmark against human creators to see if it really works.
- The Italian influencer marketing market is worth 352 million euros and growing: ignoring the topic isn't an option, but burning audience trust is worse.
53% of consumers are less likely to buy from a brand that uses AI in its social content. Yet the virtual influencer market in Europe is already worth $1.6 billion and is growing 41% a year: the question isn't whether they're coming, but whether your company is ready to use them without losing your audience's trust.
Is it really worth testing AI influencers in your company's campaigns?
It depends on your industry, your audience and how much you're willing to invest in transparency. The data shows the risk of eroding trust is high, but not the same for everyone.
An AI influencer is a virtual character, with appearance, voice and content generated or assisted by artificial intelligence, that promotes products on social media the way a human creator would. The question every marketing manager asks is simple: is it worth testing, or is it a needless risk?
The answer depends on three variables: your industry, your audience, and how well you can handle transparency. The table below will help you get your bearings right away.
| Your situation | Worth testing? | why |
|---|---|---|
| Fashion, gaming, entertainment, visuals | Yes | The audience is used to synthetic content, and lived experience matters less |
| General e-commerce with a mixed audience | Pilot first | Unknown attitudes: you need a survey or an A/B test before launch |
| Skincare, health, finance, parenting | No, for now | Recommendations require real experience: the risk of a trust collapse is high |
| Brands with a long track record of trust | Pilot first | A disclosure mistake can wipe out years of relationship with your audience |
| Startups without a loyal audience | Yes | Less to lose in terms of existing trust, and lower production costs |
What do consumers really think about AI influencers?
Most people are skeptical: nearly half wouldn't engage with an AI influencer and wouldn't trust its product recommendations.
According to a study published on ResearchGate and cited by HubSpot, 45.3% of consumers would watch an ad with an AI influencer without doing anything, 28.3% would ignore it completely, and only 26.4% would actively engage, by clicking or searching for the product.
The figure is in line with what Clutch found in its report on AI in social media: 45% of consumers say they would not knowingly engage with an AI influencer or trust its recommendation.
The emotional reaction is often immediate and final. A Reddit user cited by the source puts it plainly: “If I see something that's AI but tries to look real, it's an instant turn-off. All it takes is seeing it once from a brand, and trust is lost forever.”
There is a generational variable you shouldn't ignore, though. According to a report by LipSynthesis and MIT Sloan, consumer trust in AI avatars grew from 41% in 2022 to 73% in 2025. The share of Gen Z and Millennials who find AI videos “informative” rose from 32% to 58% over the same period. Younger audiences are more open, but they may not be your audience.
Do AI influencers build or destroy brand trust?
They destroy trust when transparency and disclosure are missing. They build it only in contexts where the audience doesn't expect lived experience.
The underlying problem is structural: an AI influencer can't actually use a product, has no genuine interests, and has no human experience to share. The appeal of human creators comes from exactly this: the credibility of someone who has tried something and tells the story.
The Clutch numbers are clear: 71% of consumers would question a human influencer's recommendation if they found out the influencer had used AI-generated content. 53% are less likely to buy from a brand known for using AI in its social content.
The risk isn't just theoretical. It also affects human creators who use AI to produce content, a topic we explored in our article on AI transparency and content marketing.
The most important finding comes from an EMARKETER/Emplifi analysis: 91% of consumers expect brands to disclose when they use AI in marketing. Only 7% say seeing it makes them trust the brand more. 31% say it makes them trust the brand less. Disclosure isn't an ethical nice-to-have: it's a business lever.
Which brands have already used virtual influencers, and what can we learn?
Several international brands have tested virtual influencers with mixed results: the most successful cases share one trait, consistency between the character and the brand's positioning.
Not all virtual influencers are AI-generated. Miquela Sousa (known as Lil Miquela) was created in 2016 using artistic CGI, traditional computer graphics, and has worked with Gucci, Prada and Calvin Klein. She was the first virtual avatar to sign with a modeling agency. Samsung created Sam, a virtual character that generated a rise in interest in Galaxy devices. The Indian brand boAt used Kyra to launch audio products.
The case most cited by the source is Aitana Lopez, created by The Clueless, a Barcelona agency: over 400,000 Instagram followers, collaborations with Amazon, Victoria's Secret, Olaplex and Razer, and reported revenue of more than $11,000 a month for the agency.
| Virtual influencer | Technology | Brand partner | Category |
|---|---|---|---|
| Aitana Lopez | Generative AI | Amazon, Victoria’s Secret, Olaplex, Razer | Fitness & lifestyle |
| Miquela Sousa | Artistic CGI | Gucci, Prada, Calvin Klein, PacSun | Fashion & music |
| Samsung Sam | Corporate CGI | Samsung | Tech & consumer electronics |
| Kyra (India) | CGI/AI | boAt | Audio & lifestyle |
The common pattern in the successful cases: the virtual character fits the brand's positioning, the target audience is used to digital and synthetic content, and the creator's virtual nature is disclosed or at least well known. None of these brands used an AI influencer to promote products that require lived experience, such as health, finance, or therapeutic nutrition.

How much do AI influencers cost compared to human creators?
Initial setup costs vary widely, but once the character is created, operating costs are significantly lower than an ongoing partnership with human creators.
The HubSpot source says that creating an AI influencer can cost anywhere from a few hundred dollars to hundreds of thousands, depending on how complex the character is. Once setup is done, though, the production costs typical of human creators disappear: locations, lighting, travel, fees for every single piece of content.
A human creator needs to sleep, eat, and take care of their own life. An AI influencer can produce content 24/7. For brands that need high volumes of content, such as e-commerce stores with many SKUs, frequent launches, or multiple markets, the operational savings can be real.
| Cost item | Human creator | AI influencer |
|---|---|---|
| Initial setup | Low (scouting, contract) | Low to high (character creation) |
| Cost per content piece | High (fees, production) | Low after setup |
| Availability | Limited (schedule, private life) | Unlimited |
| Creative control | Partial (negotiated) | Full |
| Creator reputational risk | High (human behavior) | Low (but disclosure risk) |
| Perceived authenticity | High | Low to medium (depends on the industry) |
Why everyone is talking about it now: are the Italian and European markets really accelerating?
Yes. The Italian influencer marketing market exceeds 350 million euros, and the European virtual influencer market is growing at 41% a year. Ignoring the topic is no longer an option.
The Italian influencer marketing market reached 352 million euros in 2024, up 9% on 2023, according to UPA and ONIM data presented at IAB Showcase. The 2025 estimate is around 370-375 million euros.
On the content side, ONIM (Osservatorio Nazionale Influencer Marketing) found that around 300,000 pieces of content from brand-influencer collaborations were published in 2024, 25.6% more than in 2023, with 320.7 million total interactions (+66.5%).
In Europe, the virtual influencer market generated $1.662 billion in 2024 and is expected to grow to nearly $13 billion by 2030, with a compound annual growth rate of 41.5%, according to Grand View Research.
In Italy, according to Kolsquare data, 66% of companies plan to increase their investment in influencer marketing, with an average spend of 3.37 million euros per company per year, the third highest in Europe after Germany and France. Virtual and AI influencers are no longer a curiosity: they're a budget variable.
Three real-world scenarios: how does the choice change for different Italian companies?
A fashion e-commerce store, a professional firm, and a manufacturing company have completely different needs and risks when evaluating AI influencers.
To understand where your company stands, it helps to think through typical cases. The numbers below are illustrative: they show the reasoning and are not real client data.
| Company type | Scenario | Sentiment risk | Recommendation |
|---|---|---|---|
| Fashion e-commerce (50-200 employees) | Launch a capsule line with AI influencers on Instagram and TikTok | Medium: young audience, but disclosure is mandatory | Test on a segmented audience and measure sentiment before scaling |
| Professional firm (consulting, legal, financial) | Use an AI avatar to explain services on LinkedIn | High: professional credibility depends on human authenticity | Avoid: use videos of your real team. AI can support production but shouldn't replace the face |
| B2B manufacturing company | AI influencers for virtual trade shows and product content | Low: the audience judges the product, not the creator | Feasible for demos and technical content, less so for direct sales relationships |

How do you launch an AI influencer test without burning audience trust?
You need four steps in sequence: assess the risk, build guardrails, define sentiment KPIs, and plan disclosure before you even create the character.
- 1 · Marketing managerAssess risk by categoryAsk yourself whether your product requires lived experience. Skincare, health, finance: stop. Fashion, gaming, entertainment: proceed with caution.
- 2 · Marketing + LegalBuild written guardrailsDefine in writing: allowed topics, tone, who approves each piece of content, who can block a post in an emergency, and who owns the rights to the character.
- 3 · Marketing + LegalPlan disclosure before launch91% of consumers expect transparency. The AI label must appear on the profile and on every relevant post. Check the regulatory requirements for each market and platform.
- 4 · Ads managerSet sentiment KPIs, not just reachFollowers and likes aren't enough. Measure sentiment in comments and the real conversion rate, and compare them with a benchmark of human creators on the same product.
- 5 · Marketing managerSet a stop thresholdDecide in advance at what level of negative sentiment or conversion drop you'll end the test. Without a predefined threshold, teams tend to run on too long.
- 6 · Internal teamMeasure for at least 60-90 daysThe novelty effect inflates the data in the first few weeks. Real results only emerge once the audience has had time to get used to the character.
What mistakes are most common when testing AI influencers?
The most common mistake is measuring only reach and ignoring sentiment. The second is not planning disclosure, which today is both an ethical obligation and a regulatory requirement in many markets.
Anyone working in Marketing and performance ADV knows that vanity metrics (followers, impressions, likes) don't tell you what's happening to brand trust. With AI influencers, this risk is amplified.
A second frequent mistake is treating the AI influencer as an extension of your existing creator strategy, without considering that audiences react differently. As we saw in the article on wrong AI data in campaigns, using unvalidated signals to make budget decisions is one of the most underestimated risks in paid social.
| Error | Why it is a problem | How to avoid it |
|---|---|---|
| Measuring only followers and likes | They don't detect the erosion of trust | Add sentiment tracking and real conversion rate |
| Not planning disclosure | Legal risk and a collapse of trust if discovered | AI label on the profile and on every post, approved by legal |
| Launching in high-risk categories | Skincare, health, finance: audiences don't forgive | Assess the category before creating the character |
| Test window too short | The novelty effect distorts the data | Measure for at least 60-90 days |
| No benchmark with human creators | You don't know whether AI performs better or worse | Test in parallel with a human creator on the same product |
| Full automation without supervision | No one can block problematic content in real time | Always keep a responsible person with the power to stop |
How do I know if the AI influencer campaign is really working?
You should measure at least four indicators: sentiment in comments, conversion rate, comparison with human creators, and changes in brand trust over time.
Creator strategy has changed: as we discussed in the article on creator series vs single posts, the brands getting measurable results are the ones that treat the creator as a channel with its own KPIs, not as a generic brand awareness activity. The same goes for AI influencers.
| Indicator | How it is measured | Alert threshold |
|---|---|---|
| Sentiment in comments | Manual analysis or social listening tool | More than 20% negative comments about the AI: reassess |
| Conversion rate | Click → tracked purchase or lead | Below the human creator benchmark: stop or revise |
| Brand trust (survey) | Periodic survey of the target audience | Perceived drop after launch: investigate the causes |
| Qualitative engagement | Comments mentioning the product vs comments about the character | If people only talk about the character and not the product: problem |
| Disclosure compliance | Monthly post audit | Any post without an AI label: block and fix immediately |
So, are AI influencers worth it for your company?
Worth it if your industry doesn't require lived experience (fashion, gaming, entertainment), you have a young audience used to synthetic content, you can guarantee full disclosure, and you already have sentiment KPIs in place.
A middle path: your industry is neutral and your audience is mixed. Run a survey or an A/B pilot before scaling, and set a clear stop threshold.
Not yet if your product requires credibility based on real experience (health, finance, skincare, professional consulting), or you don't yet have a system to measure your audience's sentiment.
How we see it at SHM
For us, AI is a tool that should give time and control back to companies, not replace the authentic relationship with the audience. AI influencers are a perfect example of this distinction: used well, they cut operating costs and boost scalability; used badly, they burn through years of trust built with human creators in just a few posts.
We believe the choice between human creators and AI isn't ideological, it's strategic. It should be made with real sentiment data, not enthusiasm for the technology. And it needs a measurement system that goes beyond likes, exactly as we do when we help our clients structure Meta campaigns and Google Ads campaigns with clear KPIs and comparison benchmarks.
We've worked with brands that manage large volumes of content across multiple markets: the challenge is never just producing more, but keeping consistency and trust as you scale. If you want to explore how AI can support content production without losing authenticity, the starting point is a strategic integration of AI into your existing processes, not a leap into the void. And if your team needs to understand where and how to use these tools, our AI training for companies is the right place to start.
The topic of AI influencers is also tied to how top creators are redefining their partnerships: we covered it in our article on why top creators are choosing fewer brands, a sign that the quality of the relationship matters more than the quantity of content.
Frequently Asked Questions
There is no specific law on AI influencers, but existing rules on misleading advertising and the transparency of sponsored content apply. In Europe, the Digital Services Act imposes transparency obligations on platforms, and the AI Act introduces requirements for AI systems that interact with the public. Disclosure, meaning stating that the character is artificial, is now both an ethical obligation and a concrete legal precaution.
HubSpot reports a very wide range: from a few hundred to hundreds of thousands of dollars, depending on the complexity of the character. For an SME that wants to test the format, there are more accessible solutions based on avatars generated with AI tools, but they still require investment in production, editorial management and legal oversight. The real cost also includes the time of whoever manages the character.
The most direct way is to ask them: a short survey of your current audience, or an A/B test on a small segment, gives you real data before you invest. Alternatively, analyze the comments on posts from competitors who have already used virtual influencers in your industry: sentiment in the comments is a reliable indicator of how your typical audience will react.
No, at least not in every category. Human creators bring authenticity, lived experience and a genuine relationship with their followers that a virtual character can't replicate. AI influencers can work alongside human creators for high-volume content, secondary markets or visual formats where personal experience isn't the key factor. Full replacement is not advisable, especially in categories where trust is the main purchase driver.
At a minimum: the character must be identified as virtual or AI-generated in the profile and on every relevant post. Sponsored content must be labeled as such, exactly as with human creators. Before launch, have your lawyer review the disclosure texts for every market and platform you operate in, because requirements vary.
Vanity metrics like followers, likes and impressions aren't enough. You need to measure sentiment in comments and the real conversion rate (clicks to purchase or lead), and compare this data with a benchmark of human creators on the same product. A periodic brand trust survey of your target audience completes the picture and shows you whether the campaign is eroding or building trust over time.
The real question isn't whether AI influencers work, but whether your audience is willing to trust them, and whether you're willing to truly measure it.
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.