“It would be great if there was an app that could tell you whether a student is really sick and whether they have the first signs of Covid-19. You are a teacher, not a doctor, and you have to make some kind of diagnosis,” I said to a friend during a run through Schönbusch Park in Aschaffenburg on a Sunday in November 2020. The often non-digital and cumbersome processes between schools, health departments, and doctors brought some to despair. This included my friend, the director of a school near Aschaffenburg.
That was the birth of VOIZZR.
We, software experts specializing in solutions with artificial intelligence, started researching on the internet the same day. Various universities in Great Britain, the USA, and Israel were already researching this topic. As we later found out, so were some in Germany.
When an idea grips us, we want to do it ourselves. We quickly started organizing data, building infrastructure, developing the first apps, and designing the AI algorithms. Since some of us are former Microsoft employees and have also worked a lot for Microsoft, we host on Azure in Europe.
In parallel, we entered discussions with various companies. We had never scheduled meetings with the executive boards of German retail giants or DAX companies so quickly. Due to our AI projects with some Bundesliga clubs and the DFB, the connection to sports was obvious, and numerous discussions were held. The University Hospital of Cologne wanted to pilot the solution.
After about 8 weeks, the first MVP was running.
Then everything went differently than planned!
The Corona numbers dropped in Germany, and the tests were now done by general practitioners and not centrally at the clinics anymore. So there was nothing more to test. We tested the MVP with the staff on cruise ships in northern Germany, some of which had been quarantined off Australia for months at the beginning of the Corona period. All negative. So we had a solution and didn’t know if it worked.
But the topic of voice analysis and pattern recognition had captivated us. In various projects, we had already built speech robots or analyzed media content with AI for large TV broadcasters. Week by week, new possibilities were added. New parameters.
I quickly realized that my voice is very sensitive. When I had unpleasant things to do and was under stress, I could recognize patterns. In the mornings after strenuous exercise sessions, I had significant changes. Some changes were short-term, some permanent. As humans, we often don’t notice many changes. Not because it’s not possible, but because we lack comparison or the opportunity for training. We always only have snapshots, perceive changes very poorly, and often cannot recognize obvious patterns. AI can. And now, patterns can also be recognized retrospectively, and possibly conclusions can be drawn about diseases or stress or other influences.
In sports, many injuries occur due to fatigue, increased strain, or too short regeneration. With the help of the app, phases of fatigue are now made visible in a simple way. Athletes are under massive pressure, make mistakes, and suffer psychologically. With a friendly handball Bundesliga club, we were able to identify players who were under massive pressure. The team manager then sought a quiet conversation with the player. In the winter, we were also able to identify 3 Corona cases before the normal Bundesliga testing protocol. The players were quarantined earlier and could infect others less.
New use cases are brought to the table weekly. Sports associations, clubs, and athletes virtually ‘feed’ us with ideas. In football, the topic of heading the ball and concussions has been discussed. Numerous players from England’s 1966 World Cup-winning team have died of dementia Alzheimer’s. The cause is suspected to be known. But do we really know?
My father passed away at 50 from ALS. Throughout his illness, he experienced significant changes in his voice. What if diseases could be detected as early as possible? There is no cure for ALS. Not today, but maybe in the future. What if it were possible to detect as early as possible that the CPE risk for a rugby or American football player has rapidly increased due to concussions or so-called sub-concussions? Perhaps a break could be prescribed. And all of this through the voice with the help of an app! Almost everyone has a smartphone with them at all times.
Parkinson’s also shows changes in the voice quite early on. Or even Type 2 Diabetes. We may find a cure in the future; research is working on it. Maybe we’ll reach a point where we can detect diseases very early and perhaps even prevent them with simple means. I hope so. It drives me every day. If our AI can help even a single person, it has all been worth it.
If our AI can help even a single person, it has all been worth it.
Today we have more than 50 algorithms that visualize the voice and examine it for patterns. I recognize patterns in my voice every day, maybe I will never know what they mean. That’s why we work with researchers from universities and clinics. We make our platform available here. We already have thousands of users who monitor their voice daily. From the beginning, data protection and security were important to us. We do not know the users. Many we will never meet. We want users to become more mindful. They should pay attention to changes in their body. Their voice wants to tell them something. But the responsibility continues to lie with the people, not the AI. The AI is just there to help.
Best regards,
Chris
Founder of VOIZZR