It is no secret that generative AI is helping telemedia companies produce more campaigns, creative variants and personalised promotions than ever before. But new research suggests that volume alone is not delivering better results. For value-added service providers, the real opportunity may lie in combining AI with the behavioural and cultural signals generated by online communities.
Generative AI has rapidly become part of the marketer’s toolkit. It can produce campaign concepts, social posts, video scripts, banner variations, landing-page copy and messaging content in seconds. For telemedia companies marketing games, streaming services, mobile content, subscriptions, competitions and other value-added services, that promises a significant improvement in speed and scale.
The danger is that the industry simply becomes better at producing more of the same.
New research from WARC and LIONS Advisory, conducted in partnership with TikTok, finds that 90% of marketers agree that AI has quickly become part of the creative toolkit. Some 88% say generative AI has increased the volume of creative they produce. Only 45%, however, say it has significantly improved creative quality.
That gap between output and effectiveness should matter to telemedia companies. The sector already operates in highly competitive digital environments where consumers are exposed to countless offers, subscription prompts, social posts and messages. Making it easier to create another hundred pieces of generic promotional content risks increasing noise rather than generating demand.
The brands producing the strongest creative outcomes, the research argues, are combining the power of AI with the power of community. They are not merely asking AI to make more adverts. They are feeding it richer signals about what people are searching for, discussing, sharing, remixing, watching and buying.
For telemedia, this potentially changes how value-added services are developed, positioned and marketed.
Demographics are no longer enough
One of the most striking findings is the continued reliance on basic demographic information. Some 67% of marketers say demographics remain the most common input used when prompting AI, despite 59% agreeing that traditional demographic segmentation is no longer effective, according to WARC’s Marketer’s Toolkit 2026 survey.
Only 17% of marketers always incorporate community or audience insights beyond demographics into their generative AI workflows.
This exposes a potentially serious weakness in the way AI is being deployed. Powerful creative tools are being briefed using relatively weak information. Age, gender, location and income may still offer useful context, but they reveal little about why someone joins an online community, follows a creator, downloads a game, subscribes to a video service or responds to a particular mobile offer.
Telemedia services are often built around interests, passions and moments rather than conventional demographic groups. A mobile gamer, football supporter, music fan or short-form drama viewer may have more in common with other members of that community than with people of the same age or income.
The same is true of the channels through which services are marketed. Audiences on TikTok, YouTube, Instagram, WhatsApp, RCS and other participatory platforms do not simply consume content. They comment on it, react to it, share it, remix it and discuss it with others.
These interactions provide live signals about language, humour, cultural references, emerging interests, dissatisfaction and intent. They can indicate which content formats are resonating, which creators are trusted, which offers are confusing and which themes are beginning to gain momentum.
For value-added service companies, that could be more useful than knowing that a customer fits into a particular age bracket.
From social audience to market intelligence
The research finds that 86% of marketers believe real-time audience behaviour and community signals are important to creative development. The same proportion believes these signals will have a greater influence on creative development over the next three years.
This suggests that social media’s role in telemedia marketing is evolving. It is no longer simply a place to distribute advertisements or acquire traffic. It can also provide intelligence about what products should look like, how they should be explained and which experiences audiences value.
Comments beneath creator videos may reveal that consumers do not understand how a subscription works. Online discussions may show that users want shorter-form content, different payment options or greater control over renewal. Search behaviour may indicate growing demand for a particular game, entertainment format or digital utility before that demand becomes obvious through sales data.
These signals can then be fed into AI systems to shape the creative brief. Marcos Angelides, Managing Director of L’Oréal Lab and Head of AI Operations at Publicis Media, puts the issue plainly: “Where the advantage comes is the data that you use to train those models, because AI is only as good as the data it’s trained on. You’ve got to have behavioural data. You’ve got to know what people actually do, not just what they say they do.”
For telemedia providers, that behavioural data could come from a broad range of consented sources. It may include content engagement, subscription activity, messaging responses, search behaviour, app usage, customer-support questions and the conversations taking place around services on participatory platforms.
The opportunity is not to use AI to watch individuals more closely. It is to identify aggregated signals that reveal what communities collectively value and how their interests are changing.
Creating an intelligence loop
The WARC, LIONS Advisory and TikTok report goes on to describe this process as an “Intelligence Loop”: a continuous cycle in which community participation and AI-powered creative reinforce one another.
The cycle begins when participation generates signals. People search, comment, share, create and buy, producing layers of information about culture, interest and intent.
Those signals can reveal demand. They may identify an emerging trend, content format or audience need before it becomes explicit in conventional market research or campaign reporting.
Demand then shapes the creative. Community signals are fed into the AI briefing process, changing the information from which the system produces concepts, messaging and content.
Finally, the creative fuels further participation. Material that feels relevant to a community is more likely to generate engagement, conversation and sharing, producing new signals that inform the next campaign.
The significance for telemedia lies in the cumulative nature of the process. Each campaign should make the next one more informed. A subscription entertainment provider might identify a growing conversation around a particular genre, use AI to develop and test relevant campaign concepts, monitor the response and then use those findings to refine not only its advertising but also its content selection and service design.
A gaming provider could analyse community discussions and creator content to understand which mechanics, characters or rewards are attracting attention. AI could help translate those signals into different campaign executions, while performance and community responses generate the next round of intelligence.
A messaging or mobile engagement provider could use aggregated response patterns to understand which language, timing and formats consumers find useful rather than intrusive.
This is fundamentally different from using generative AI as a production shortcut. It turns creative development into a live learning system.
Creators are not just media inventory
The research also challenges companies to reconsider how they work with creators. Creators have frequently been treated as distribution channels: individuals with an audience who can be paid to promote a product. The new model positions them as sources of intelligence.
Creators are often deeply embedded within specific communities. They understand their language, jokes, concerns and unwritten rules. They can see which ideas are gaining traction and which brand interventions are likely to be rejected as inauthentic.
For telemedia brands, particularly those selling entertainment, games, competitions, lifestyle content and subscription services, this knowledge can be highly valuable. AI-enabled systems may help identify creators based not merely on follower numbers but on community fit, cultural relevance and campaign objectives. They may also help providers analyse the subjects, formats and audience responses surrounding creator content.
That does not mean automating the creator relationship or attempting to manufacture authenticity. It means using AI to help companies understand where genuine alignment exists before they begin producing and distributing campaign material.
