The future of audio AI for media: Highlights from our INMA keynote
On October 24, 2025, my co-founder Patrick O’Flaherty and I shared our insights into the future of audio AI for media at INMA’s Media Tech and AI Week in San Francisco.
Here are four highlights from our keynote speech, which focused on AI article narration:
1. Brands will sound as distinct as they look
As audiences spend more time with audio and video content, how your publication sounds will become just as vital to its identity as how it looks.
Voice plays the central role in sonic branding. Just as a person’s voice tells us so much about who they are and how they feel, an article’s voice influences how listeners perceive your publication and connect with your story.
AI voice cloning makes it possible to use unique, on-brand voices at scale.
For example, we helped News Corp Australia clone the voices of dozens of journalists in newsrooms across Australia, including legendary cricket commentator Robert 'Crash' Craddock.
Sonic logos, background music, and sound effects can also make audio articles more distinctive and immersive. This helps audiences to recognize your content across platforms and form lasting connections with your brand.
2. Multimodal consumption will drive multimodal journalism
As multimodal news consumption grows, publishers must craft stories that work fluidly across formats.
Journalists and editors will increasingly consider how their work sounds as well as how it reads, writing in ways that feel more conversational. They may also favor topics that lend themselves well to spoken formats, such as human-interest stories.
AI will help to bridge the gap between formats, automatically adapting the base content for optimal listening, viewing, or reading experiences. For example, by generating spoken summaries of visual charts.
These advancements will give journalists the freedom to focus on their preferred mediums while maximizing their stories' impact and reach.
3. Audio will become integral to product design
Audio is too often an afterthought in product design—something bolted onto text-first interfaces rather than built in from the start.
But that’s changing.
The future lies in audio experiences that are intrinsic to product design, treated as a core part of the user experience rather than an optional add-on layer.
Some publishers are already experimenting with audio-centric design.
Earlier this year, Jysk Fynske Medier (JFM) increased listen rates tenfold by implementing BeyondWords’ playback-by-paragraph feature, which lets readers start listening by clicking anywhere in the text.
With our new persistent playback feature, available through our player SDKs, listeners can keep listening as they navigate between pages on a website or app.
4. Hyper-personalized audio will deepen engagement
In the future, AI audio will adapt to each listener in real time.
The voice’s tone, accent, and pacing will automatically shift to suit personal preferences or inferred listening habits, deepening listener engagement.
For example, narration might sound brighter and more energetic during morning commutes, then softer and slower during late-night listening sessions.
One of the clearest opportunities for personalization lies in regional accents, which can be tailored to a listener’s preferences or location. This will make audio more engaging and easier to understand.
Publishers like The Irish Times are already using regional accents to strengthen connections with localized audiences.
Prepare for the future of AI audio
Audio production and consumption is changing fast, and media companies must adapt to stay ahead.
With BeyondWords, you can automatically generate, distribute, and monetize audio versions of your articles using advanced AI voices or voice clones. To learn what we could do for your publication, book a meeting with our team.
