Opinion: With AI, MAM systems can finally deliver on their potential

Traditional media asset management (MAM) systems hold a crucial function. They organise and store digital assets in searchable repositories, however, many legacy MAM systems are built around static metadata tagging, demanding extensive manual input – and often falling short when scaling for today’s vast content demands. This manual approach is both labour-intensive and limits the system’s ability to generate contextual inferences within archives.

Multimodal AI offers a new way of working by automating the creation of quality indexing through vector embedding or commonly used data formats that can be integrated into a MAM database. This empowers MAM systems to recognise relationships between assets and radically improves searchability and content discoverability through the lens of concepts, rather than just keywords. Integrating AI with traditional MAM infrastructure can leverage advanced capabilities like semantic search, summaries, and sound bites, finally realising MAM systems’ original vision as a flexible repository in a content-heavy world.

Legacy MAM systems are often ill-equipped to handle the massive metadata volumes AI generates. Simply layering AI onto older systems only taps a fraction of the data’s potential. To fully harness AI’s transformative power, organisations must adopt platforms designed for the scale and complexity of AI-driven workflows.

Enabling new content discovery workflows

Multimodal and generative AI are not only reshaping media management but opening doors to workflows previously beyond human reach. By processing and integrating data types such as text, images, and audio, AI can produce comprehensive summaries and detailed scene descriptions that previously required massive human and financial resources.

For example, 24/7 Arabic news service Asharq News applies multimodal AI to newly broadcast content, automatically generating transcripts, facial recognition data, and summaries that enable its producers to stay aligned with — and even ahead of — current trends.

Another significant use case lies in archive digitisation. As vast magnetic tape collections, once the backbone of media archives, rapidly reach ‘end-of-life’, there is now a daunting need for digitisation. Manually cataloguing each tape would require a substantial time investment, which is prohibitive for many organisations. One of Moments Lab’s US TV broadcast customers calculated that it would take 100 years to index 200,000 hours of archive. Multimodal and generative AI changes the game, not just in relation to the speed at which it can index digitised content. Advanced AI models are showing enormous potential in identifying what’s inside physical tapes before digitisation, providing insights into the value of that content.

Speed of development calls for constant re-evaluation

The next generation of AI has brought an unprecedented speed of technological change. Companies operating with advanced AI models must continuously reassess their applications and workflows to ensure they remain competitive. This evolution, which is showing to be faster than even the shift to HD, requires service providers to take a dynamic approach to managing the role of AI in media operations. AI-driven content discovery platforms equipped with built-in evaluation at the core and supported by dedicated research teams will push technology providers towards proactive innovation in AI video understanding. This will also intensify the need to keep pace with rapid advancements and maintain rigorous output quality assessments — a challenge that will grow as AI continues to evolve.

Identifying inefficiencies in media workflows

Effective AI implementation calls for thinking in terms of outcomes rather than outputs. The first step is recognising specific inefficiencies within a media organisation’s workflows. Low-quality metadata and fragmented indexing are frequent issues in traditional MAM systems, limiting their usefulness for rapid search and retrieval. Identifying these pain points allows organisations to strategically apply AI to create tangible improvements in media operations. This approach, centred on specific use cases, helps focus AI’s potential on areas where it can deliver the most value. The result is not just the integration of AI but a true transformation of how content is managed, accessed, and used.

The integration of multimodal and generative AI into MAM and content discovery workflows represents a significant step forward in the efficiency and effectiveness of media operations. To fully realise AI’s benefits, the media industry needs to confront and adapt to the inherent challenges of traditional MAM systems – specifically the way they store and search through data. Embracing this dynamic landscape will allow media professionals to meet growing industry demands with better agility, depth, and innovation.