Acoustic AI: Should we be listening in to improve performance?

Forget the background music at your favourite cafe. There is absolutely no way you can go for a coffee in London (and I’m sure most cities) without hearing the murmurs from different tables…

“So, how are you using AI? How do think it will affect our sector?”

There is no question that the noise around AI has reached fever pitch. And the stakes are high: The generative AI market has already produced 13 unicorns (reaching a $1B+ valuation), taking an average of 3.6 years, around half the average time of a unicorn to reach this status!

I’ve spoken to many different entrepreneurs over the last 6 months, looking to place their bet in this rapidly evolving space, alongside early-stage investors who are opening up their investment theses to accommodate.

However with daily news, updates and debates, navigating generative AI space currently feels like white water rafting (like Meryl Streep negotiating down those treacherous rapids in ‘The River Wild’ – I still get sweaty feet watching this…)

But let’s take a step back from the rapids to something more serene…

24 hour context

Away from the noise, there are some very interesting other forms of AI growing in prominence, especially relevant to human performance.

Working within sport we are often trying to understand what is happening with our athletes over 24 hours.

In the high performance training centres where we work, everything very well controlled and monitored (training load, rehabilitation, nutrition), but what happens in the ~20 hours athletes’ are away from the training centre, which is a crucial time when their mind and body, recover and adapt to the training stress? 

Is there any additional background information which could easy be captured, which would help us manage their health and performance each day?

Working within elite sport we are often trying to understand what is happening with our athletes over 24 hours.

Time to listen up?

As I mentioned in my initial article of this series, I’m really interested in how our monitoring tools (within sport or other high performance settings) can monitor our athletes/ talent, with minimal daily burden…

(E.g. not giving biological samples, taking photos, adding data into an app), to capture background data on those who might less motivated (to be clear: this is with the users express consent). 

On this, last week I listened to a really interesting podcast conversation between Jason Calacanis and Hyfe founder, Joe Brew, about acoustic AI and this new health tech startup.

Hyfe is aiming to help patients and doctors solve for the root cause of the common cough, affecting 10-20% of Americans chronically.

It uses acoustic AI for digital cough monitoring, to identify and time stamp coughs from a microphone (on a smart phone or watch) over a period of time (e.g. 24 hours) to build an accurate history.

Data can also be correlated with data from other streams (nearby weather conditions, heart rate, etc.). 

Use cases within performance?

Within the health and performance space, we’ve recently seen ‘chronic’ monitoring tools such as continuous glucose monitoring (CGM) and heart rate variability (HRV).

This got me thinking about the potential use cases for information-rich acoustic AI. Some initial thoughts include:

Voice monitoring for fatigue detection within sport, music or business settings. Voice as marker to detect sleep deprivation, or even to sense vocal fatigue in those use their voice as a key part of their role (e.g. coaches, musicians, politicians)…

More widely, and of interest coaches, healthcare practitioners, or anyone who delivers large amounts of verbal information each day to clients; the ability to capture large volumes of audio (e.g. a years worth of consultations/coaching sessions) could perhaps be used to AI train models, to optimise the coaching/consultation process?

As with much of the rapidly evolving AI industry, this presents ethical issues around user consent and data privacy (an important issue which we’ll continue to return to).

We know that the ‘quantified self’ market (users ability to self-monitor) will continue to grow year-over-year.

My bet is that we’ve only just started capturing some very rich, low burden, information which could really help with both; 1) athlete/talent monitoring and 2) the consultation/coaching process.

If you’re interested in this or building something in the space, get in touch!

Keeping it shorter today…Shipping the work, as Seth Godin would say. More to follow on this….

James