Björn Ühss is one of the co-founders of AudioStack, the leading AI audio production platform for global brands, publishers and advertisers. The takeaways in this list are adapted from a speech he delivered at Radio Days Europe 2025.

You can listen to an AI summary of this takeaways list here as a podcast:

AI transformation in audio is happening at a pace far quicker than many of us expected - whether we are ‘in the know’ or not.

By now, audio that’s produced by traditional production processes is functionally indistinguishable from audio produced by AI.

For consumers, there’s no difference. For creatives and advertisers, it makes all the difference. 

Whether you create content at scale, care more about personal experience, work in an agency  or work in audio sales, you need to understand how this landscape has changed.

This blog will tell you everything you need to know.

Key Takeaway 1: Without AI, even the most streamlined audio production processes are defunct.

If you are involved in audio production, you know four essential steps go into producing an audio asset like a broadcast ad.

• Writing the script

• Hiring a voice actor

• Mixing and mastering

• Exporting/Delivering the final product (e.g. into your AdTech platform).

These aren’t just inefficient for the reasons you know, such as the time it takes to plan and approve scripts, or mixing and mastering audio manually.

They also require separate, specific and non-transferable expertise (your scriptwriter is rarely also competent as a producer and vice versa); and if at any stage (even late in production) you decide on a stylistic change, the intensive production process starts again.

AI-driven audio is changing not only each step, but making the workflow as a whole quicker and more flexible.

Key Takeaway 2: AI Audio Production is doing for audio what Canva did for visuals

With AI, Audio Production for ads can be as simple, fast and accessible as creating a visual asset on Canva. This means:

• Writing script using generative AI (with nothing more than a product description if you need).

• Choosing a natural, lifelike voice to read it with the right delivery.

• Delegating mixing and mastering (reducing noise, eliminating echo, compressing it) to an AI designed for the task.

• Versioning it in seconds, making ads as personalized as our social media feeds.

It doesn’t just transform how the process happens, either. That changes who can do it. Where your previous process needed a writer, voice talent, a producer and an AdOps expert, these can all be a single person.

Key Takeaway 3: Speech and audio aren’t the same thing: that’s where AI makes the biggest difference

At the pace AI audio production capability has moved, you can be forgiven for thinking that it’s enough to automate only your voice acting with AI, using an AI text-to-speech voice.

These are, after all, incredibly convincing in cadence, pitch, tone, intonation and emotion.

However, what you may not know is that AI audio production is already beyond capable of mastering every other element of production. Not just in the steps that make audio sound ‘clean’, but in creative elements, like selecting the right music for the mood based on your prompt, and fading it in and out at the right times for the biggest impact.

This skill isn’t the reserve of specialists anymore. AI has democratized it.

Key Takeaway 4: AI Audio is becoming the rule, not the exception

If 2024 was about exploration and adoption of AI in the advertising space, 2025 is about proliferation. 

Already, the largest global brands, agencies, and AdTech providers are producing audio ads entirely with AI for a number of different use cases.

For example, a global news provider, Newscorp transformed news headlines into broadcast-ready radio ads with AI, whilst a national retailer, PENNY, created and versioned hyperlocalized, dynamic creative ads for each of its supermarkets in Germany in a fraction of the time, and for a fraction of the cost of traditional DCO advertising.

For brands, publishers and the creative/technical platforms that support them, the workflow and creative optimizations AI audio platforms bring have already proven their worth. 

The early adoption stage for AI-powered audio has well and truly passed into the majority adoption stage.

Key Takeaway 5: AI ad production is transforming how sellers engage buyers.

We already know that AI audio production platforms democratize audio ad creation. This isn’t just game changing when work is being delivered. It changes how we pitch to, engage and win new business.

Why?

Because salespeople can now be empowered to do what was only previously possible for audio production experts..

Now, thanks to AI, self-service audio production platforms like AudioStack enable sellers to create their own personalized audio demos for pitches. This lets them:

• Create more engaging and impactful customer campaign presentations

• Win more pitches by producing more compelling assets at a greater volume

• Drive revenue by winning pitches and reducing the cost of producing each pitch

Ads are created in just a few clicks, ready in minutes, and sounds as professional as if they were created by professionals in a studio - all with the ability to instantly iterate them at scale.

• Fact: sellers can also provide more value to their customers by providing the creative for free if higher media budgets are booked, or by providing various versions of an audio creative to allow A/B testing and performance optimization.

In short, this is revolutionizing traditionally long and risky sales cycles of outreach into something much more interactive and engaging, and it shines nowhere better than live at sales conferences.

Key Takeaway 6: AI is enabling unparalleled flexibility for creators of audio spots

Take a listen to this concept, conversational ad for Nespresso.

🎧 Listen here

Let’s divide it into different parts.

• You may recognize George Clooney - his voice isn’t cloned; it is the standard, original voice, embedded into the rest of the innovative, AI-extended ad.

• The end of the ad (“visit Nespresso.com…”) was a lifelike human voice generated by AI-driven text-to-speech.

The female voice in the ad was synthesised with AI. But the voice actor was actually male - in fact, it was the same person who produced the ad.

Speech-to-speech (STS) capability (in this case, voice recording and transformation with AI) allowed the creator to record his lines, with the desired intonation and delivery, and output a lifelike female voice.

For comparison, a tool like Siri was developed with hundreds of hours of voice. If you want to produce something more lifelike to produce for your next ad, you need just 30-40 minutes of voice.

Key Takeaway 7 - AI has changed Ad Operations and dynamic creative optimization for good

AdOps and AdTech specialists that deal in data-driven creative know that their rewards aren’t easily won.

Ads work better when they are personalized and relevant to each listener - be it calling out a listener’s location, or (in a more technically-demanding scenario) inserting a relevant ad into a podcast by genre.

But creating, QA’ing and uploading these into different line items into your AdTech platform is a mammoth task.

And it’s one that AI has made far easier. From a creative standpoint, the production of 100s to 1000s of ads, ready for national distribution, is cut from weeks into minutes - with QA and upload assisted by intelligent algorithms.

As regards ensuring these campaigns work, AI is changing the way iteration is handled. Now, you can ensure the results you receive from the first run of the campaign can be retooled and redeployed as new audio assets at incredible speed and efficiency, thanks to automation and optimised workflows that make no difference between 1, 10, 100 or 1000 ad creatives.

Just see how this worked for the Women’s Rugby World Cup in 2025: 

Key Takeaway 8: Audio Production has forever changed - leadership & change management hasn’t.

Digital transformation as a concept always sounds exciting. And in fact, adopting a new, AI-driven platform into existing ways of working is less intimidating in reality than in theory.

But the only blocker to it, as you will know from your experience, is managing people’s appetites for (and reactions to) change.

In this respect, AI audio is no exception.

If we have a parting message for those who will ultimately be responsible for signing off on the next big change, it’s this: AI requires a new-thinking. One of adaptability, fast adoption and willingness to experiment. 

And if you don’t know where to start, you can ask yourself a few simple questions to get on the right track:

1. What kinds of efficiency, organisational results and competitive advantage can I drive by using AI and automation?

2. How can we experiment early, appoint change makers and accelerate the learning curve for AI within our company?

3. What tools and technologies can we leverage to drastically improve or disrupt our core business models?

4. Who in my organisation has shown early signs of being a change adopter and can help be a catalyst for positive change?

Fortunately, if your role makes you feel like you are expected to be an expert in everything, you don’t - especially when it comes to AI.

Seek expert support. Test the tools that work. Run pilots earlier to validate them within your organization. And above all, map the AI solution to your core business model to ensure it will do what you want it to.

With AI Audio, hearing is believing

Using AI goes far beyond having new, more efficient ways of generating voices for your ads. 

AI will help you deliver 10x more efficient end-to-end audio production processes; make the audio sales process more interactive than ever before; and unlock personalization for your dynamic creative campaigns at a scale never before seen.

Want to see how that can transform the way your brand, agency or platform produces audio? You won't believe what's possible until you see it for yourself.

A seismic transformation is taking place. If you need to know where to start; and how to take advantage, you can read our whitepaper, in collaboration with AWS and the University of Cambridge, "GETTING AI TRANSFORMATION RIGHT" below: