Audio may be the oldest of the so-called “new media,” but it’s also true to say that it’s currently getting the most bang for the buck through AI. If you’re producing podcasts or radio programs in 2025, artificial intelligence tools can significantly improve both your workflow efficiency and the quality of your final product.
Transforming Transcription and Editing
AI first started to make its impression on audio through the rapid improvement of transcription tools. One of the first transcript tools I used was called Descript, and it’s fast become an essential one-stop shop for audio producers (although I still opt to complete my work in a traditional DAW). While Descript is principally aimed at the video market, the platform can be effectively used for audio-only productions.
The first way it can transform your current workflow is through automatic transcription. When you import your raw audio file, it will create a transcript that can be edited like a document. If you highlight and cut a section of the transcript you don’t want in the finished recording, it will cut the corresponding audio. The cut is precise and sonically invisible. For podcast makers with a print background, this will feel like a revolution.
As an old radio hand, I use this function to remove the unessential chit chat at the top and tail of my files, though I still do finer edits on a traditional DAW. However, scanning a fairly accurate transcript is an excellent way to get a shape of the work needed before getting your hands dirty with the audio itself.
Removing Filler Words
Another tool that has become part of my workflow is Descript’s ability to remove filler words from both the transcript and the audio file. This feature defaults to a long list of words, so it’s best to select only the specific words you want to remove to preserve the natural feel of the conversation.
I often find this tool essential because you can set it to remove “umms” and repeated words. On a radio program I recently produced with two very chatty hosts, it removed over three minutes of “umms,” meaning more of the actual discussion could make it into the final program. There’s certainly a case for leaving in some hesitations—if you ask someone a curly question and they “umm” at the start, that’s an important tell—but 99-100% of these fillers are just our brains hitting the pause button, and they can kill the flow of the conversation.
Using tools like these to remove filler words requires sensitivity, but they can save a tremendous amount of time and improve the final audio quality. During tight turnarounds, this can be a lifesaver.
Enhancing Audio Quality
A number of new audio tools are emerging that process and enhance recorded sound. These AI tools use technology similar to what Peter Jackson employed on “Get Back” to enhance the mono recordings from the late 1960s.
The enhancement tool I regularly use is called Studio Sound in Descript. This simple tool does two key things: it removes the noise floor and reduces the room sound from a recording. The change this tool can bring is often dramatic. The thin, hollow sound of an original recording is removed, replaced by a full voice with a quiet background.
If your guest connects remotely via Zoom or similar platforms, the improvement to their audio quality can be remarkable. To the untrained ear, it can make remote participants sound like they’re in the studio with you. Currently, this processing isn’t done in real time, so it’s only useful for pre-recorded material. While I’ve used the word “remarkable,” this tool does have its limits. Studio Sound has one control (0-100%), and it’s best used subtly as the audio can develop an over-processed sound at higher settings. In my experience, around 60-65% typically yields positive results without artificial artefacts.
Alternative Platforms
Before this turns into an advertisement for Descript, I should mention that other audio and video platforms used by many podcasters offer similar capabilities. Riverside, for example, includes their own audio enhancement tool called Magic Audio, which I’ve found very useful and similar in scope to Descript’s Studio Sound.
Other notable options include Adobe’s Enhance Speech feature in Premiere Pro and Audition, and Auphonic’s automated audio post-production system, which has been a staple for many podcasters for years.
Conclusion
There’s much more to say about these emerging tools and platforms, but if your current workflow has remained unchanged over the past five to ten years, you’ll almost certainly benefit from adding some of these AI-powered solutions to your toolkit. They not only save time but can also significantly improve the quality of your productions—allowing you to focus more on content and less on technical cleanup.
Whether you’re producing a radio show, a podcast, or an independent audio documentary, these tools can help you work more efficiently while delivering a more professional-sounding final product.
About the Author:
Anthony Dockrill is a Digital Producer at Pod Jam and the former Program Director of 2SER FM Sydney.
Contact Anthony at [email protected]