Prime Focus Technologies Transforms Content Localization with Agentic AI on AWS

Prime Focus Technologies (PFT) provides cloud-based AI and media services to broadcasters, studios, and streaming platforms. Its flagship product, CLEAR, supports complex media workflows by managing and monetizing content at scale. With 16 years in the market, CLEAR currently manages over 14 million media assets for global customers—including Disney Star, CBS, Lionsgate, and Channel 4—and has helped these organizations reduce their media operations costs by 30 to 40 percent through automating processes.

Building on the success of CLEAR, PFT began developing CLEAR AI, an agentic AI–powered platform to automate more complex tasks. Muralidhar Sridhar, senior vice president and global head of AI at Prime Focus Technologies, explains, “Since the launch of CLEAR, we’ve seen media operations evolve from manual to software-driven asset management, but AI is taking that to a new level—automating human-in-the-loop workflows entirely for greater efficiency.”

As part of CLEAR AI’s development, PFT sought to improve the speed and performance of its AI agents for long-duration tasks such as subtitling, language translation, and transcript generation—while keeping operational costs under control. This would require a highly scalable, managed generative AI service. “Our strategy is always to use best-of-breed solutions, and we wanted an approach that could support the platforms our customers already use,” says Sridhar.

After testing several leading solutions that failed to deliver the right balance between speed and cost, PFT turned to Amazon Web Services (AWS) to support development of its agentic AI platform. The team selected Amazon Bedrock, a fully managed generative AI service, to build and deploy AI agents more efficiently. To accelerate development and ensure alignment with best practices, PFT engaged the AWS Prototyping and Cloud Engineering (PACE) team, which collaborates with customers to rapidly prototype and build solutions on AWS. Working with the PACE team, PFT adopted Anthropic’s Claude large language model (LLM) plus LangGraph, which transforms LLMs into agentic, reasoning-based systems. “AWS PACE is so well-versed in Amazon Bedrock that it immediately becomes an asset to your team,” recalls Sridhar.

Next, PFT wanted to integrate Amazon Bedrock running Claude with its internally developed LangGraph API to support agentic workflows. “The PACE team came back with a solution that we could use to start creating AI agents for content localization tasks,” says Sridhar. PFT built a supervisor architecture in which multiple AI agents collaborate on long-duration localization workflows. “We achieved the balance of speed and cost for our AI agents that we wanted using Amazon Bedrock,” Sridhar recalls. “With the AWS PACE team supporting our software engineering team, we accelerated our time-to-market, completing work on our agents in weeks rather than months.”

Having upskilled on Amazon Bedrock, PFT began developing a service for customers to enhance live sports events. The goal was for LLMs to generate game tags and highlights during cricket matches automatically. When PFT experimented with LLMs through external APIs, it was found that latency could be quite high for live tagging, but the figure dropped dramatically with generative AI on AWS. Says Prathap Simha, senior technical architect at Prime Focus Technologies, “Using Amazon Bedrock and Anthropic’s Claude, our generative AI solution can create tags and highlights of match events close to real time, providing customers with a service that engages viewers.”

Broadcasters, studios, streaming platforms, and other PFT customers have already reduced content localization costs by 20–30 percent and accelerated content localization by 30–40 percent by using AI agents on the CLEAR AI platform. These agents also deliver up to 20–30 percent greater accuracy in subtitling, language translation, and transcript generation than other vendor solutions. “Our AI agents on Amazon Bedrock are not only helping automate media workflows but also reviewing and improving processes over time to deliver increasing value to our customers,” says Sridhar.

With deep industry expertise and a modern agentic AI framework, PFT is reinforcing its leadership in the media sector. “Embedding intelligent agents within the CLEAR AI platform is a significant milestone—and one we’re building on with the support of AWS,” says Sridhar. “Running CLEAR AI on AWS empowers creative and operational teams with AI agents that can reason, scale, and integrate securely within enterprise ecosystems.”