BERLINALE 2025: The Danish consultancy firm's Rikke Flodin and Lars Tversted presented their new tools, showing how they can help creatives develop successful projects

l-r: Rikke Flodin and Lars Tversted during the talk

On 14 February the European Film Market hosted a panel titled "Using AI-Backed Anthropology to Build Strong Audience Awareness", which saw the participation of two reps of Publikum, a Danish consultancy firm. Publikum’s Head of Audience Insights, Rikke Flodin, discussed how the foundation of Publikum stemmed from a fundamental lack of understanding of the concept of “audience” within the Danish film industry. She summarised this issue by stating, “Data people struggle to inspire creative people, and creative people struggle to convince data people that their ideas are truly valuable.” She also explained that the project was born out of a desire to bridge these two perspectives, enabling them to work together more effectively.

Flodin went on to describe their approach as a fusion of technology and human insight, encapsulated in the term “AI-enhanced anthropology”. This methodology has already been applied to around 100 projects, ranging from animation to documentaries and commercials. Notable examples include Flee [+see also:
film review
trailer
interview: Jonas Poher Rasmussen
film profile], Sweet Dreams [+see also:
film review
trailer
interview: Ena Sendijarević
film profile], and Sick of Myself [+see also:
film review
trailer
interview: Kristine Kujath Thorp
interview: Kristoffer Borgli
film profile].

The presentation continued with a decalogue of recurring themes observed in audience reactions to films. The first point highlighted was the statement: “Coming-of-age stories often lack true resonance with young people.” Young audiences feel that these films should have a lighter tone. Too often, they are overly harsh and fail to offer a fresh perspective beyond everyday teenage life, ultimately disengaging the very audience they aim to attract. When writing a coming-of-age story, the key question should be: “What lighter elements could make this film appealing to young people, rather than just arthouse audiences?”

Secondly, the findings indicate that audiences are more invested in relationship-building than in identity development. This implies a need to focus on mutual, dynamic relationships rather than one-sided narratives. Another key observation was that “genre awareness surpasses genre fandom”. Flodin advised filmmakers to clearly define the three main genres that describe their project or, if it is an original genre blend, to be fully aware of its unique positioning.

Publikum’s research also reinforces the importance of remembering that films are, at their core, a form of entertainment. Audiences often anticipate plot developments because they consume a vast amount of content. The challenge for filmmakers is to ask: “Can we give audiences what they expect without resorting to clichés?”

Additional insights included the idea that a film’s build-up does not always need to occur at the beginning. This means ensuring that the opening sequence has the right pacing and avoiding overly explanatory scenes, as audiences enjoy piecing together narratives themselves.

Another crucial takeaway is that audiences talk about and remember characters more than plots, themes, or stories. They also prefer "character explosions" over gradual "character development". Viewers are drawn to characters who evolve in unexpected ways, as this adds both intrigue and emotional connection.

Finally, the research suggests that European audiences are ready to be challenged. The real risk often lies in not pushing boundaries far enough, rather than going too far. Stories, scenes, characters, visuals, language, and dilemmas should be explored even more boldly to create lasting impact.

Flodin described Publikum’s approach, stating, “We don’t use old scripts to predict potential success; instead, we analyse current conversations. We also don’t focus on box office performance but rather on sentiment and emotions. We dislike automation and prioritise collaboration, meaning the AI never works independently.”

Their methodology consists of three key steps: Vision, Zeitgeist, and Narrative. First, they analyse the project’s materials. Then, with the assistance of an algorithm, they examine how people discuss a particular theme online, drawing from over 100,000 online data sources that are clustered based on meaning. Finally, the Narrative phase involves testing audience reactions, gathering feedback from over 500 participants on specific elements of the project.

This approach helps answer critical questions such as: “Which character generates the most discussion? How does the audience perceive the balance between humour and drama? How can we attract two very different age groups to the same film? How can a film resonate beyond the traditional arthouse audience and reach a broader viewership?”

Finally, Lars Tversted, a partner at Publikum, introduced “Logline”, a new predictive AI tool based on score predictions and IMDb reviews of approximately 60,000 European films. During workshops, participants develop three different pitches, testing variations of the same story with distinct key themes. The tool then identifies which films in the IMDb database resemble each logline. "Logline" has the potential to cross markets and expand a film’s reach into territories that may not have been considered initially.