From intuition to intelligence: How Nordic AI is reshaping filmmaking decisions

Across the Nordic audiovisual ecosystem, artificial intelligence is increasingly less about automation for its own sake and more about sharpening creative and strategic decision-making. Two Nordic firms - Publikum and Vionlabs - illustrate how AI can operate at very different moments of the filmmaking lifecycle, yet share a common ambition: to help stories travel further by understanding audiences more deeply, without flattening creative intent.

Publikum: bringing the audience in before it’s too late

Publikum, founded as a Copenhagen-based audience-intelligence consultancy rather than a pure software provider, positions itself firmly at the development end of the pipeline. Its core proposition is to move audience insight upstream - before scripts are locked, casts finalised or shoots completed - while the cost of creative misalignment is still avoidable. As co-founder and CEO Niels Alberg explains to Nordisk Film & TV Fond, Publikum’s “combination of hard and soft data helps align a project’s stakeholders around a more coherent shared vision - from screenwriter to sales agent - creating a more unified and frictionless conversation and outlook on both the project’s artistic ambitions and its commercial strategies”.

Traditional research tools, Publikum argues, arrive too late. Test screenings and focus groups react to finished or near-finished works; Publikum instead asks what “emotional and cultural signals are already present in the audience before the project exists as a finished object”. Its methodology combines three pillars: cultural relevance, narrative feedback, and cinematic expectations.

Cultural relevance is mapped through AI-supported zeitgeist analysis, drawing on up to 100,000 data points per territory to explore how audiences emotionally relate to a film’s themes. Rather than starting from marketability, the analysis addresses “the societal and cultural dimensions before the commercial ones”, based on the belief that emotional depth is itself a key commercial driver.

Narrative feedback, meanwhile, is gathered through anthropological methods that prioritise radically honest, individual responses, unshaped by group dynamics or dominant voices - often surfacing lived experiences that filmmakers can directly translate into creative adjustments. The third layer, cinematic expectations, is delivered through Publikum’s PlotBounce tool, trained on 165,000 curated films, allowing teams to test how a logline may be culturally positioned and compared - iteratively - across development, production and launch.

Importantly, Publikum does not promise deterministic answers. Its philosophy is built around what it calls “structured reflections” and “creatively acceptable predictions”, rejecting black-box certainty in favour of transparent dialogue. Data and audience quotes are shared openly with producers and funders, enabling qualitative discussion rather than prescriptive conclusions. This approach has tangibly shaped projects, from fine-tuning character traits and genre balance to enabling segmented marketing strategies - exemplified by Leonardo Van Dijl’s Julie Keeps Quiet (Julie zwijgt), where distinct audience groups were addressed with differentiated messaging at launch.

This emphasis on decision-support rather than automation extends to Publikum’s institutional collaborations. Its new pilot programme with Germany’s MOIN Film Fund invites companies to work across entire slates rather than single titles, aligning individual projects with long-term strategic positioning. With a newly opened Hamburg office and locally hired staff, Publikum is explicitly scaling its Nordic model into more complex funding landscapes while remaining people-centred rather than purely platform-driven.

Vionlabs: turning story and emotion into scalable content intelligence

If Publikum’s work concentrates on helping films become what they need to be, Vionlabs focuses on ensuring that finished content is discovered, experienced and monetised in ways that respect story and emotion. Built as a media-native AI company, Vionlabs analyses video at scene level across image, audio and text to understand mood, pacing and narrative structure.

As explained by the Vionlabs team - Marcus Bergström, Leona Unrath and Arash Pendari - the company’s mission is “to help the media and entertainment industry turn video into actionable intelligence, so that every viewer gets a more relevant, emotional and effortless viewing experience”. Unlike general-purpose AI solutions, its proprietary multimodal technology is designed “to understand mood, pacing, narrative structure and visual/emotional cues”, allowing platforms to move beyond title- and genre-based logic.

This approach underpins two complementary product families. The Discovery Suite, including Editorial Lab and Creative Lab, supports editorial, marketing and creative teams with enriched metadata, mood-based recommendations, thumbnails and preview clips, while the Efficiency Suite automates operational processes such as intro and credit detection, ad-break creation and contextual advertising.

Vionlabs stresses that its core differentiation lies in being “not a general-purpose AI company, but a domain-specific, media-native AI company built exclusively for video and storytelling”, with models built and continuously trained in-house on media-specific data from more than 100,000 titles. The impact is already measurable at scale: A major global streaming service recorded a “+17% increase in viewing minutes across titles using AI-generated previews” and “+44% more titles added to watchlists”, while a large European TV operator integrating mood-based recommendations achieved a “+15% increase in started streams” and an “85% user preference for mood-matched recommendations in A/B testing”.

Together, Publikum and Vionlabs illustrate a distinct approach to AI in filmmaking - one that privileges transparency, cultural specificity and emotional understanding over automation hype. Whether shaping a film’s creative DNA early on or ensuring that finished stories reach audiences in the right emotional context, both firms point to a future where data does not replace intuition, but sharpens it.