SVG Sit-Down: Brahma AIs Patrio Cummins on the Possibilities of AI (and the Misconceptions)

AI was a big buzz on the IBC show floor but there weren't a ton of companies who were leveraging it in meaningful and mature ways. One that was, however, was Brahma AI, formerly known as Prime Focus Technologies. Patricio Cummins, Brahma AI, SVP and Global Head of Technology Sales, sat down with SVG Global Executive Director Ken Kerschbaumer to discuss Brahma AI's newest offerings and enhancements as well as the key trends impacting the industry's move to AI as a key component of production workflows and operations.

SVG: What is new at the show this year? What is your big story regarding new technology and products?

Cummins: We have been investing heavily in AI for seven to eight years, and we are glad it has become a major focus in the industry. Over a year ago, we launched Brahma AI Studio, which handles all our AI-led content creation solutions like ATMAN for digital humans, or Vaani for content localization. While our product Brahma AI Core serves as our Media Asset Management (MAM) and workflow orchestration layer for managing and distributing content, Studio is strictly focused on production and content creation. We use it to generate high fidelity digital humans, voice localization, promo and trailer creation, and generating mini-shows or dramas from long-form content for highly specific audiences. All of this is AI-based.

SVG: How does Studio work in practice? Are you bringing live camera feeds directly into the system, or does it ingest existing video files?

Patricio Cummins, Brahma AI, SVP and Global Head of Technology SalesCummins: Studio handles production, but not live production. It works with pre-recorded, file-based material as source content but generates new content out of it. We use this technology across multiple verticals, including drama, unscripted content, and sports. Speaking of sports, our newest development is with a major North American sports leagues using our MAM technology. They are using our MAM technology for highlights and content distribution.

SVG: In general terms, how do you see AI helping with MAM operations?

Cummins: It is all about automation. Previously, you needed a large team looking after a MAM, but now many of those functions are AI-led. That doesn't mean AI does it all; we still recommend having a human in the loop for added value and quality control. AI is not replacing creativity, but it excels at time-consuming tasks and helps businesses scale rapidly.

For example, we provided an AI-powered compliance solution for a large company in the UK. Preparing a 30-minute episode for compliance used to take a human operator 20 to 22 minutes. With our solution, it takes three minutes with a human in the loop to review the output. That efficiency scales massively across a catalog of thousands of hours.

SVG: Going back to Brahma AI Studio, once you ingest a file, what are the actual processes? Does the AI work on simultaneous workflows?

Cummins: It is multi-layered because we deploy different AI agents. One agent might only analyze audio, while another looks at images or video, depending on what you are trying to achieve. Once a file is ingested and the proxies are extracted-we work mostly with proxies-you can run a number of concurrent operations. In a single pass, the system can create a highlights reel, generate social media clips, write taglines for publication, draft written summaries, and extract keywords or metadata for enrichment so the file goes to storage fully tagged and ready to be utilized in a number of downstream activities.

SVG: When speaking with customers or prospects, do you encounter any misconceptions or fears about AI and how to properly approach it?

Cummins: Yes, I can think of three common concerns. First, the concern of AI replacing humans. In our world, at least, humans are irreplaceable. This is why we always recommend a human-in-the-loop workflow. AI is here to assist and augment, not replace. We call this Mind2, the fusion of human intelligence and machine intelligence where humans set intention and AI powers execution.

Then there is a valid concern because cloud AI costs can escalate quickly if you process everything indiscriminately. We get around this by optimizing spend across the entire operation using our proprietary machine wisdom technology. Instead of running AI through an entire hour of content-which would send costs through the roof-we analyze chunks intelligently. We do a first pass with audio to establish context. Then, if a shot remains static, we only need to analyze one representative frame. If your face is prevalent in a file, we only need to recognize it once.

And then, lastly, there is the I can do it myself' mentality. Many companies with internal teams think they can build these tools in-house. However, once you look at the total cost of owning and operating this over 12 to 18 months-not just implementing it, but keeping it running smoothly-bringing in an external vendor makes a lot of sense. We invest heavily in R&D, secure patents, and deliver proven products that provide real value.

SVG: AI speeds up processes so much that it opens up entirely new workflows. Instead of spending hours logging or struggling to find the right clip, teams now have the freedom and efficiency to create more content and drive new revenue. It seems like it changes the entire dynamic of media asset management.

Cummins: Absolutely, and that brings up another major shift. Because we now have an AI layer sitting on top of the MAM, users can interact with it using natural language, conversationally, just like they would with ChatGPT. Historically, only technical staff accessed the MAM to search and index based on metadata. Now, anyone-a social media coordinator, an editor, a producer, or a QC operator-can query the MAM and act on content directly, within their permissions.

A real-world example

LINK:	https://www.sportsvideo.org/2026/09/30/svg-sit-down-brahma-ais-patrio-...
	See more stories from svg

Most recent headlines