Artificial intelligence is changing how companies build products and manage their operations. But adopting AI is not just about adding new tools. It also requires companies to rethink their workflows, team structures, and the way people work. At Harman International, this approach extends to audio product development, where AI could help simplify engineering, bring advanced features to more affordable devices, and make listening experiences more personal. The company is also exploring how audio systems can understand users’ preferences and surroundings and adjust settings automatically, with little manual input.
In an interaction with Digit at JBL’s 80th anniversary event in Seoul, South Korea, Armin Prommersberger, CTO of Harman International, spoke about how the company is adopting AI, its potential across consumer, professional, and automotive audio, and how it could make audio experiences more seamless.
Q. You have argued that companies need to rethink their systems and processes to adopt AI effectively. What does this mean in practice at Harman?
Armin Prommersberger: At Harman, we have a company-wide initiative focused on accelerating AI adoption. The first step is to understand how people work before deciding where AI can help.
A workflow involves more than the steps recorded in a company’s software. Employees may have to call someone for information, export data, perform a manual task and then import the data again. We map these steps to understand the complete process and why it exists. Only then do we redesign it with the people involved and identify where AI can help.
Software development is a good example. Earlier, a team might have 20 programmers, a team leader, a project manager and a scrum master working together to solve a problem. But the team could face fragmented code repositories, missing documentation and a lack of shared context.
AI can help us build software much faster, but the quality of the outcome depends on having the right information, a well-organised codebase and a reliable source of truth. We are therefore rethinking team structures and bringing together people with different skills, including system architects, product managers and engineers who know how to use AI effectively.
The change is significant. Teams can now make multiple code commits and submit several pull requests within an hour. But this pace can also be mentally draining. We have introduced mandatory coffee breaks every 90 minutes to help people step away from their screens.
The biggest challenge is not necessarily the technology itself. It is the human side of change. You need to understand existing workflows, what technology makes possible and what people are ultimately trying to achieve. Simply adding AI everywhere will not make a process better.
Q. How is AI changing the way Harman develops audio products and improves the listening experience?
Armin Prommersberger: AI has two broad applications for us: enterprise AI, which helps us work more efficiently, and product AI, which helps us develop better products and experiences.
On the enterprise side, AI allows us to test features faster, explore new designs and make changes more consistently. It helps us move through development cycles more quickly.
On the product side, there are visible AI features and others that work in the background. A clear example is the JBL BandBox range, which uses AI-powered technology to separate musical elements. As a drummer, I find this useful because I can play along with my favourite songs while removing the drum track.
But I am also interested in AI that works without requiring users to interact with it directly. Audio systems can use information from microphones, speakers, sensors and the content being played to understand the situation and adjust the experience.
Imagine that you are watching a film on a Saturday night. An intelligent system could recognise your preferences, prepare the surround sound system with suitable Dolby Atmos settings, switch the TV to its movie mode and dim the lights. You would only need to ask it to play the film.
We could also bring the knowledge of our audio engineers into a virtual system that lives in your headphones or speakers. Rather than sending an engineer to every customer’s home to adjust the sound, the system could continuously optimise audio settings for the best possible experience.
Q. Advanced audio features have traditionally been associated with expensive equipment. Can AI make these capabilities more accessible?
Armin Prommersberger: I firmly believe AI is already making many capabilities more accessible. We are getting better at synthesising and distilling models so that they serve specific use cases without requiring enormous computing resources.
A model with billions of parameters may not be practical for a pair of true wireless headphones. These devices have limited memory, processing power and battery capacity. We therefore need to refine models so they can run efficiently on the hardware available.
This is where companies such as Harman and Samsung have an advantage. We understand the constraints of consumer devices and can develop AI capabilities around those limits.
Once we can run a feature on a product such as the JBL Tour Pro, we can explore bringing similar capabilities to other products, including speakers. The JBL BandBox is another example. Capabilities that previously required professional music production software can now be accessed through a consumer product that costs a few hundred dollars.
This does not mean every advanced feature will immediately appear in every product. Models still need to be adapted to each device’s hardware and intended use. But as we become better at making them smaller and more efficient, more people can access capabilities that were once limited to expensive professional equipment.
Q. Harman operates across consumer audio, professional audio and automotive systems. Which of these areas stands to benefit most from AI?
Armin Prommersberger: I think the impact will be significant across all three, although the applications will differ.
In professional audio, consider a large concert venue using hundreds of loudspeakers and a complex mixing system. A small group of highly trained engineers must manage the system and ensure that the sound remains consistent across the venue. Much of their work involves maintaining settings and managing the equipment rather than directly improving the artist’s performance.
AI could automate some of these routine tasks, monitor the condition of individual loudspeakers and help protect them when necessary. This would allow sound engineers to focus more on the quality of the performance.
In consumer audio, the opportunity is to create experiences that users may not have expected from everyday products. Headphones could continuously optimise their sound, while home audio systems could work more naturally with TVs and other devices. Users should not have to read lengthy manuals or navigate multiple settings to get started.
Automotive audio presents a different challenge. A car is a noisy environment, and road noise can affect the listening experience. AI could help maintain consistent sound quality without requiring drivers to adjust equaliser settings or other controls while driving.
In each case, the objective is to reduce the effort required from the user or the professional operating the system. These are not merely possibilities for the distant future. We are already seeing relevant capabilities emerge in products and systems today.
Q. What will the future of AI-powered audio look like when the technology works as intended?
Armin Prommersberger: For me, the future is about removing friction. We can talk about advanced features all day, but if a system does not work when you need it, the experience becomes frustrating.
I want to interact naturally with my devices without having to remember which buttons to press, which app to open or which settings to change. I should be able to say, “Please play AC/DC in the living room,” and the system should take care of the rest.
The experience could become even more personal. As a drummer, I might ask the system to play AC/DC while removing the drum track so that I can play along. A capability that was once limited to a specific product could eventually become available across the home audio ecosystem.
This is where intelligent orchestration becomes important. AI could understand the user, the devices available, the environment and the content being played, then coordinate everything in the background.
I am not particularly fond of buzzwords, but I do think we are moving towards a world where technology becomes ambient. You may not always need to look at your phone to interact with an AI assistant. You could speak to your JBL speaker, which connects to your AI and helps you complete a task.
The technical side should become almost invisible. You should not have to think about how the system works. You should simply be able to use it naturally. That, to me, is where the real value of AI lies.
For more such content, keep reading Digit.in.
Also Read: Best portable Bluetooth party speakers under Rs 30000 to buy in this festive sale