AI Innovations for VFX and animation

By Michael Knott and Michael MacLaren

AI promises to revolutionize the VFX (visual affects) and animation industry, but AI is not new to this industry. While practical generative AI use cases are still emerging, the true value and cost remain uncertain. We evaluated potential use cases to determine the most promising opportunities. From automated rotoscoping to storyboarding, our findings reveal significant potential for time savings, cost reduction and impressive ROI (return on investment).

The VFX and animation industry are at the forefront of enhancing visual experiences in film, television, advertising, gaming and immersive entertainment. Picture your favorite scene in a recent movie, and it's almost certain that some form of visual effects artistry were involved. Indeed, VFX is used in almost every film, TV show and game made today.

AI and machine learning (ML) are already integrated into many aspects of production and post-production, and the industry is now poised for a new stage of AI-driven optimization and transformation. But the question is, which applications in VFX and animation stand to benefit most from the exciting new possibilities ahead?

Roland Berger carried out a study exploring various AI use cases, using a cost - benefit analysis framework to highlight their potential impact. This article highlights its key findings. In an industry often dazzled by cutting-edge technology, it offers practical, measurable outcomes rather than theoretical possibilities.

What we did: our cost-benefit analysis

AI technologies offer immense value to the VFX and animation industry through substantial cost and time efficiencies, for example, by automating content creation, enhancing realism and enabling collaborative tools. While options for "quick win" and budget-friendly AI use cases are currently limited, there are numerous high-impact opportunities for pipeline enhancement.

In our analysis, we evaluated 6 specific use cases across several AI technology types, from computer vision to deep learning and natural language processing (NLP). Each was considered in terms of benefit/value factors (efficiency gains, quality improvements, scalability etc.) and cost factors (setup, OPEX, etc.). These factors are outlined in the following two graphics.

Results: use cases rated by value and cost

The value-cost analysis allowed us to map the use cases according to their overall potential. This matrix indicates that VFX and animation houses can capture maximum value by focusing on computer vision AI use cases, including auto-rotoscoping, object and scene deletion, and image upscaling and enhancement.

A key reason for the higher potential of computer vision use cases is that they are relatively low-cost while still offering high value. This makes them a strategic entry point into AI . The technology is mature, with numerous pre-existing libraries and tools (for example, OpenCV and TensorFlow), speeding up implementation and integration. In turn, this lowers risks as known parameters make the investment more predictable, while rapid rollout reduces the chance of the technology becoming obsolete.

Below we outline each of the use cases and their cost and value rationale.

Case study: automated rotoscoping

Automated AI rotoscoping is one of the most promising short-term applications of this technology, offering a clear example of how we apply our cost-and-value framework. By significantly streamlining the traditionally labor-intensive process of manual rotoscoping—tracing live-action footage frame by frame to produce animated sequences—AI introduces substantial efficiency gains. However, the extent of these time savings depends heavily on the type of film being produced. Different genres involve varying levels of complexity, which influences both the workload and the value of automation.

To demonstrate the potential impact, we evaluated five film genres—action, period, sci-fi, horror, and romantic comedies—using a structured approach. The rotoscoping workflow was divided into three phases: initial keyframing (focusing on primary frames), high-detail refinement (addressing secondary frames), and low-detail interpolation (handling in-between frames). For each genre, we assessed how AI’s capabilities—such as automated mask generation, enhanced motion tracking, and improved frame interpolation—could reduce manual effort at each stage.

Our analysis revealed that time savings range from 20–65% depending on the genre, driven by factors such as scene complexity, camera movements, and the presence of effects like motion blur and CG explosions. Of course, no shot or film is created equally when it comes to VFX requirements, however our analysis showcases meaningful efficiency gains, allowing resources to be reallocated to higher-value tasks. The following pages provide detailed time-saving estimates for each genre, along with concrete examples of how AI delivers these improvements. Our estimates use a standard eight-hour workday, but we know the reality of this industry often involves longer days.

A framework for success: strategic steps for AI integration

So how can VFX and animation players effectively harness the opportunities presented by such promising AI technologies as automated rotoscoping? A strategic roadmap is essential. Traditional industry software vendors – e.g., Autodesk, Foundry, SideFX – are all making investments in AI/ML enhancements to their software packages. Hence, it is critical for VFX/animation players to understand vendor roadmaps and make strategic choices in terms of vendors and buy vs build.

Roland Berger has developed a comprehensive AI use case framework – raise – which can support any VFX and animation studio to develop AI foundations, scale deployments and achieve AI excellence. It is based on the following key strategic steps:

1. Budgets: Outline a dedicated budget to determine project resourcing and scale
2. Data and systems: Assess and upgrade pipeline systems to enable the use of internal data for AI use cases
3. Compliance and risk: Establish and agree on clear compliance frameworks regarding the use of training data based on intellectual property (IP) ownership
4. Strategic prioritization: Prioritize use cases based on their value to the business within the defined budget
5. Piloting: Start with low cost technologies to demonstrate value and build a case for further investment. Use financial metrics like IRR and ROI to evaluate the potential value of AI technologies
6. Maintenance and supervision: Continuously evaluate and manage the newly implemented AI technologies , and look for areas of model improvement and enhancement
7. Ethical standards and governance processes: Establish guidelines for ethical AI operations, including transparency and explain ability and set robust data management policies
8. Change management and partner coordination: Develop strategies to address internal resistance to AI tools and to manage the external partner ecosystem effectively

A clear way forward

With so many options to explore, the careful evaluation of the cost-benefit of new AI technologies will be critical to the future success of VFX and animation players. Evaluation will help to determine AI readiness, find the best fit for existing technology stack and decide on whether to take an early adopter or more cautious approach.

Regardless of which route is chosen, we strongly recommend following a structured framework, such as Roland Berger's eight-point plan, to ensure a balanced and effective AI implementation. Our detailed roadmap helps to avoid common pitfalls and maximize return on investment. It ensures studios can tailor AI adoption to their specific needs and maintain control over costs and outcomes, while complying with complex AI-related IP legislation/management. By adhering to a responsible approach to AI, studios can ensure they not only enhance their creative capabilities but also safeguard their long-term viability in a rapidly evolving landscape.

For more information about our analysis or AI use cases in the VFX and animation industry, please contact one of our experts.