The recurring theme in this column is that change is the only constant in media technology. For the past three years, generative AI has dominated almost every conversation in our industry.
However, something has shifted in the last year. According to Gartner’s 2025 “Hype Cycle for Artificial Intelligence,” generative AI has officially entered what analysts call the “Trough of Disillusionment,” that phase where inflated expectations give way to implementation realities. As you prepare for next month’s NAB Show in Las Vegas, this is the lens through which you should view every demo, every booth and every pitch.
As we all know, AI is not new to professional media. Machine-learning algorithms have been embedded in our workflows for well over a decade—in recommendation engines, in content fingerprinting and rights-management systems, in automated quality control and in speech recognition.
Over the last few years, generative models captured the public imagination and suddenly every product and every booth had “AI” stamped on it. The time has come to separate proven tools from science projects and AI veneers on other technologies.
As you walk the show floor this year—including the AI Innovation Pavilion and the expanded Creator Lab with its dedicated AI sessions—your goal should be clear this time: focus on the technologies with a demonstrated track record of solving real problems in professional media workflows.
Where AI Is Already Delivering
The good news is that there are many areas of professional media where AI-driven tools have moved well beyond the pilot stage and are delivering measurable value today.
Accessibility and localization represent perhaps the most mature and impactful deployment of AI in our industry. AI-powered captioning, transcription, subtitling and audio description have reached a level of accuracy and scale that was unexpected just a few years ago. During major global sporting events, AI captioning systems are delivering live subtitles across thousands of hours of simultaneous coverage.
AI-powered translation and dubbing services are enabling broadcasters to reach multilingual audiences in near-real time, and these capabilities are now being deployed to help entities meet new requirements in accessibility. This year’s NAB Show will feature numerous exhibitors demonstrating these capabilities that deserve your serious evaluation because they represent proven, deployed technology.
Media asset management and metadata generation is another area where AI is an essential tool. The broadcasting industry produces staggering volumes of content, and traditional manual tagging processes cannot keep pace.
AI-powered metadata systems can automatically extract visual elements, generate semantic descriptions, identify audio components, create temporal markers and apply consistent taxonomies across entire content libraries. We have seen large reductions in manual cataloging time after deploying AI-driven metadata automation, along with dramatic improvements in content discoverability.
The good news is that there are many areas of professional media where AI-driven tools have moved well beyond the pilot stage and are delivering measurable value today.”
Beyond these two pillars, look for proven deployments in automated quality control and compliance monitoring, where AI can flag lip-sync issues, subtitle overlap and standards violations. Postproduction automation tools that can isolate audio stems, identify natural ad-break points and generate multiple content variations are also moving into production environments.
AI-driven ad segmentation technologies that analyze content to find optimal insertion points are creating real revenue for broadcasters willing to adopt them. AI is even solid in some areas of content production of non-AI content, such as camera tracking and rotoscoping.
Avoiding the 95% Failure Rate
No matter how hard you try, you won’t be able to avoid some hype and a few demos that look cutting-edge. What is important now is to avoid wasting time or money on projects that will not provide real value in a reasonable time frame.
A landmark 2025 study from MIT’s NANDA initiative: “The GenAI Divide: State of AI in Business 2025,” analyzed over 300 AI initiatives, conducted 52 organizational interviews and surveyed 153 senior leaders. The finding all of us should look at closely is that 95% of enterprise AI pilots delivered zero measurable return on investment. While 80% of organizations explored AI tools and 60% evaluated enterprise solutions, only 5% reached production with measurable business impact.
Critically, the MIT researchers found that this is not a failure of technology but of execution. The AI models themselves are typically powerful, but the failures stem from what the study calls the “learning gap”—enterprise deployments strip away context, feedback and adaptability, leaving static tools where dynamic systems are needed. Users need tools that learn from feedback and can be customized to fit into existing workflows.
So how do you, as a media technology professional, evaluate new and less-proven AI opportunities at the show without becoming part of that 95%? Here is what the research—and decades of experience with technology transformation in our industry—tells us works:
- Start with a specific workflow problem, not a technology. The most successful AI deployments in media began with a clearly defined pain point—a bottleneck, a backlog, a staffing challenge—and then found AI solutions that addressed it. Do not buy a solution in search of a problem.
- Pilot narrow, then scale. Midmarket firms in the MIT study scaled successful AI pilots in 90 days, compared to nine months for large enterprises. The difference was scope: smaller, focused pilots with clear success metrics outperform ambitious enterprise-wide rollouts every time. Avoid the big bang. Be agile.
- Ensure existing workflow integration. Ask vendors how their tool fits into your existing media technology environment. If they cannot answer that question concretely, think carefully about how to approach it or not.
- Blend human expertise with AI capability. Research from multiple sources indicates that human-AI pairing boosts productivity. The goal is augmentation, not replacement. In professional media, where editorial judgment, creative instinct and regulatory compliance matter enormously, this is important.
- Measure what matters. Do not chase vanity metrics or vendor benchmarks. Define your own success criteria—time saved in postproduction, accuracy rates in captioning, reduction in manual metadata tagging, improvement in content discovery times—and hold vendors accountable to them.
If this advice sounds familiar, it should. Every major technology transformation I have witnessed in this industry over the past several decades—from tape to file, from SDI to IP, from on-premises to cloud—has followed the same pattern. The technologies that endured were the ones that solved real problems, integrated into existing operations and delivered measurable value. AI is no different.
While You’re There…
Last year, our team at Deloitte gave a presentation titled “Future Unscripted: How to Be Ready for Anything in an Uncertain Media Landscape,” in which we talked about resiliency to ongoing change. This year we will be doing a session on practical AI— “Make AI Make Sense”—where we will dive deeper on the principles discussed here. Please feel free to check it out and join in the dialogue.
As you plan your time at NAB Show, bring your skepticism and your curiosity in equal measure. The hype cycle has crested. What remains is the hard, rewarding work of making AI actually make sense in your facility, your workflow and your business. That is where the real opportunity lies.
