Particle6 launches Xicoia, an AI talent studio | what it means for AI filmmaking

Particle6 founder Eline van der Velden has introduced Xicoia, a studio focused on building and managing hyperreal digital performers for screen projects, games, brands, and social video. The announcement came during the Zurich Film Festival in September 2025, with the studio's first public performer, Tilly Norwood, presented to industry attendees.

The pitch is simple in its logic. Treat virtual performers as talent, not as one off effects, and give them creative direction, schedules, and business models that look familiar to producers and agencies.

What Xicoia Is Building

Xicoia describes its digital performers as having biographies, voices, and continuing arcs that can appear across film and television, social channels, and game engines. The framing pushes the conversation from tools to outcomes. A virtual performer with a throughline can headline a promotional campaign today and return for a limited series tomorrow, without retraining the audience each time.

Tilly Norwood is the first of these performers to be publicly presented. Norwood is a fully developed virtual actor with an established visual identity. The Zurich Film Festival presentation drew interest from industry attendees, with Deadline reporting the announcement alongside trade coverage from Televisual.

The business model Xicoia describes is closer to a talent agency than a software company. Performers are managed assets with defined creative identities, not configurable avatars that producers customize from scratch for each project.

The Access Argument

For filmmakers working outside major production hubs, virtual performers reduce one of the practical barriers to production, specifically talent availability. A creator producing content in a location where on screen talent is scarce, expensive, or difficult to schedule can work with a digital performer whose appearance is consistent across every production pass.

Iteration also improves in a meaningful way. Cycling through accent adjustments, wardrobe looks, and performance notes for a live actor requires booking windows and pickups. For a digital performer, those changes require generation passes, not shooting days. Early development work moves faster when creative choices can be tested before any crew is contracted.

The third argument is ownership clarity. A digital performer created under clearly defined terms has explicit licensing scope, territory rights, and usage limitations. By contrast, contracts for human talent often leave gaps around what happens to the performance across different distribution contexts and future derivative works.

How to Direct a Virtual Performer

For filmmakers, the practical question is how you direct a performer whose presence is software and stagecraft rather than a body on set. The answer requires building the same directorial precision you would bring to any production phase into the generation process itself.

Where a director gives notes to a live actor between takes, with a digital performer those notes become prompt specifications, reference images, and audio guides before generation. The direction happens upstream rather than on set. This requires translation skills, specifically the ability to convert a performance note into a technical specification that the generation pipeline can act on.

Some directors find this more clarifying than the live set process. The requirement to specify intent precisely before seeing results forces decisions that might otherwise be left to improvisation. Others find it constraining because the feedback loop between a live actor and a director is replaced by a generation cycle with a longer turnaround.

The most practical approach is to use digital performers for scenes where the direction can be specified precisely upfront and where multiple generation passes are affordable. Scenes that depend heavily on real time creative chemistry, where the best takes emerge from live interaction, remain better suited to human talent.

The Rights Infrastructure Required

The IP ownership structure around virtual performers is not yet standardized. When a studio creates a digital performer and uses it in a production, the performance IP, the likeness IP, and the underlying model IP are three separate layers with different owners, different licensing terms, and different obligations.

Xicoia's model attempts to handle this at the studio level by owning and managing the digital performers as defined assets. For productions licensing Xicoia performers, the expectation is that the licensing terms cover the performance IP and likeness IP together, similar to how a talent agency clears a human actor for a specific production.

Any production using virtual performers should establish explicit terms for training sources, allowed uses, territories, and downstream derivatives such as trailers, games, and merchandising. These terms should be as specific as any principal deal, not a general technology license.

This three layer structure requires more careful documentation than traditional talent agreements. A human performer's contract covers the performance and, separately, any agreed use of their likeness. With a digital performer, the model IP adds a third dimension that has no equivalent in standard talent contract templates.

Early productions that establish clear precedents for how these layers are licensed will define the frameworks that become standard. Legal teams with early experience in virtual performer licensing will have an advantage when those frameworks solidify across the industry.

Production Workflow Implications

A production that integrates Xicoia managed digital performers would plan differently from one casting human actors. There are no availability conflicts, no travel requirements, and no scheduling constraints tied to talent calendars. The performer shows up as consistently as the pipeline allows.

However, the production still requires a director who can specify what the performer should do and a technical team that can execute the generation pipeline to deliver consistent quality on schedule. The human creative labor does not disappear; it shifts from set management to specification and quality review.

For brands and commercial productions, the consistency benefit is particularly strong. A digital performer can maintain identical appearance across years of campaign materials, which is difficult to guarantee with human talent across aging, hair changes, and other factors that affect visual consistency in long running campaigns.

For narrative productions, the tradeoff is different. Digital performers free up budget that would go to scheduling and availability, but they require investment in the generation pipeline, quality review infrastructure, and the technical team that operates both. A production team without prior experience in AI generation pipelines will face a steeper learning curve that offsets some of the schedule advantages in the near term.

International productions benefit from the global availability of virtual performers in a specific way. A digital performer who appears in a production shot in one country can be licensed for reshoots or additional content produced in a different country without any of the travel, work permit, or residency complications that would apply to human talent. For productions that span multiple countries and production periods, this removes a category of logistics problem entirely.

Tilly Norwood and What the Debut Signals

Tilly Norwood is the first publicly presented Xicoia performer. The details released around the Zurich Film Festival announcement describe her as a fully developed digital actor with an established visual identity, backstory, and voice rather than a configurable template.

The decision to introduce a named, specific character rather than a range of customizable options signals Xicoia's intent to operate in the performer space rather than the avatar tool space. A named performer with a defined identity can accumulate audience recognition across projects. An avatar that studios customize per project does not.

For producers evaluating the announcement, the relevant questions are what rights package accompanies a Xicoia performer license, what production contexts are covered, and what the exclusivity terms look like. A performer with a defined public identity that appears in multiple productions simultaneously creates brand consistency questions that a traditional talent deal would handle with clear exclusivity provisions.

The Zurich audience for the announcement represents one of the industry's most concentrated groups of producers and financiers with independent project mandates. Reaching that group with an early stage tool announcement is a deliberate strategic move. The industry's gatekeepers evaluate projects at major festivals, and placing a new virtual talent concept in that context positions it as a production resource rather than a tech demo.

The Xicoia Approach vs. One Off Effects

Virtual performers have appeared in productions for years without anything like Xicoia's stated model. Visual effects studios have produced digital doubles for stunt sequences, marketing materials, and specific production scenes on a project by project basis. What distinguishes those one off effects from what Xicoia proposes is continuity.

A digital double created for a stunt scene in one film does not carry over to a sequel or a franchise spin off without a new negotiation and a new build. It exists as a production asset tied to a specific project, not as a performer with an ongoing career. Each new appearance requires starting the creative development process over.

Xicoia's proposition is to maintain a library of developed performers whose visual identity, voice, and creative persona are stable assets that can be licensed across projects without rebuilding from scratch each time. The investment in character development happens once and amortizes across multiple productions. For franchise content and brand campaigns that require a consistent face over years, this is a meaningfully different value proposition than project specific effects work.

Whether audiences will accept these digital performers in narrative roles rather than as supporting or background characters is a separate question from whether the technology produces convincing output. Audience acceptance depends on emotional engagement with the character, and that depends on writing, direction, and editing at least as much as visual fidelity.

What to Do on Your First Project

Start with paperwork and proof of concept before committing to a digital performer for a full production. Add a clear rights rider that covers training sources, model licenses, allowed uses, territories, and downstream derivatives. Build a short internal pilot to measure audience acceptance, editorial overhead, and the real cost of finishing.

Credit virtual performers explicitly in end cards and marketing material. Label AI contributions so partners and viewers understand what they are seeing. This protects the production from later authenticity questions and builds a transparent record for broadcasters or distributors who require disclosure.

Keep a session log for every generation run, including reference inputs, prompts, seeds, model versions, and review notes. The ability to reproduce a specific performance months later depends entirely on that documentation. Assign a human lead for performance direction and final cut approval so there is a clear creative decision maker at every stage of the pipeline.

For independent creators, platforms that monetize content can create unexpected obstacles. Automated systems that misclassify AI generated content may affect distribution access. TikTok's AI moderation system wrongly removed independent narrative films from monetization in May 2026, demonstrating that disclosure and platform compliance are practical production concerns, not just ethical ones.

Where Virtual Talent Fits in the Broader Industry

Xicoia launched in a period when the question of AI performers was moving from experiment to industry conversation. Several studios and agencies were evaluating digital talent as a supplement to human casting, particularly for global campaigns requiring consistent multilingual versions and for productions where talent availability is a structural constraint rather than a scheduling problem.

The Zurich Film Festival venue for the announcement reflects an attempt to position Xicoia within the film industry conversation rather than the technology industry. Film festival audiences include producers, distributors, and financiers who evaluate projects, not tool evaluators who benchmark software.

Whether virtual performers become a mainstream production tool depends on audience acceptance, which is the variable that neither technology nor legal frameworks can determine in advance. If viewers believe the performance, the pipeline becomes a production option. If they do not, the technical capability remains a novelty regardless of how advanced the underlying generation becomes.

The productions that answer the audience acceptance question most clearly will be those where the digital performer carries dramatic weight in scenes the audience cares about. Background and support roles do not test the technology in the way that matters commercially. A leading role in a story that audiences follow emotionally is the real test, and it requires the full combination of writing, performance quality, and directorial skill that any leading role demands.

Xicoia's Zurich launch positions the studio ahead of the inflection point when those tests become common. Being established in the market when producers are actively looking for virtual performers, rather than entering after the pattern is set, is the strategic logic behind an early public launch at a premier industry event. The work now is building the performer library, the rights framework, and the industry relationships that convert early positioning into actual production credits.

The AI FILMS Studio video workspace provides tools for generating AI video footage and character animation for productions that want to test creative concepts with AI generated visuals before committing to a full pipeline decision.

The most important early signal for Xicoia will be which production credits the studio accumulates in its first year. A virtual performer in a short film, a brand campaign, or a streaming series pilot gives the industry a concrete example to evaluate. Those early credits answer the audience acceptance question with actual audience data rather than industry speculation.

Van der Velden's background in building digital entertainment businesses positions Particle6 as a company with commercial and creative experience rather than a pure technology startup. That distinction matters to producers evaluating whether to include a virtual performer in a production with real budget at stake.

---

Sources

Deadline | Televisual