Earzz is a British startup behind a Health Tech World Award-winning monitoring technology aiming to advance safety, independence and care outcomes.
The company was founded by CEO Pradyumna Thiruvenkatanathan who has been combining acoustic engineering and AI for around 20 years.
Thiruvenkatanathan’s innovations include a gravity sensor and a tool that enables BP to ‘listen’ to oil wells to identify what’s inside them.
Thiruvenkatanatha told Health Tech World: “Earzz was something more personal, inspired by grandmother, who didn’t receive the best care in her later years.
“I wanted to use my skills to improve social care.”
Earzz harnesses AI to track vocalisation trends such as shouting and coughing so behaviour, such as sundowning in dementia patients, can be detected in dementia patients and flagged in real time.
The tech was developed with two communities in mind: Deaf and hard of hearing people, to help them to understand and ‘read’ their environment through visual and haptic alerts, and care homes, particularly those dealing with fall risks.
Thiruvenkatanathan says: “In many care settings, outdated tools like pressure mats are still being used.
“These mats beep when someone steps on them, but they have serious limitations, especially for people with dementia.
“The mats are often dark and dementia patients can mistake them for holes, so they avoid stepping on them, and this leads to falls. Plus, the mats are wired, adding more tripping hazards.”
The Earzz system, on the other hand, sits passively in the room, listening for trigger sounds that can alert carers, even before the resident gets out of bed.
The team wanted to move away from the baby monitor model, where someone has to actively listen for trigger sound, taking up valuable care resources..
The AI build into the Earzz device processes and interprets sounds on its own, only alerting carers when necessary, freeing them to work on actual caregiving, rather than monitoring.
Pradyumna Thiruvenkatanathan
Thiruvenkatanathan says: “The tech is especially impactful in dementia care.
“Traditional nurse-call systems rely on the resident pressing a button—something many dementia patients can’t do.
“Yet, every dementia care bed still comes with one. That’s not just outdated—it’s unfair.
“Our system supports personalised care.
“For example, if a person isn’t at high risk for falls, you might only need to monitor for distress or vocalisation. The AI tailors alerts to each individual.”
This personalised approach is seeing impressive results.
A recent case study showed an 87 per cent reduction in falls. And two care homes that recently deployed the tech saw zero falls in over three months.
The development of the technology was not without its challenges.
Thiruvenkatanathan says: “While I had 17–18 years of experience on the algorithm side, building compliant hardware was a whole different challenge.
“We had to meet rigorous regulatory requirements, go through certification and ensure everything was fit for purpose.
“Then came perfecting the AI. It’s not just audio relaying; the system had to understand what it was hearing—coughs, screams, movement.
“We needed the AI to distinguish between a resident getting out of bed versus a carer walking past the room.”
The team also had to build all the backend infrastructure from the ground up, including its sound recognition engine which can learn new sounds incredibly quickly.
Thiruvenkatanathan says: “For instance, a deaf mother reached out about detecting an infusion pump alarm for her son.
“Within two days, we trained the system to recognise that sound.
“That experience led to a collaboration with one of the world’s largest medical device manufacturers.”
Earzz has a number of other collaborations in the pipeline, but the main short term goal is deepening their presence in the care home space.
So what would Thiruvenkatanathan advise others looking to innovate in this space?
He says: “Build tech that’s easily adoptable.
“It’s not just about innovation—it’s about integration.
“Care homes already have established practices, so your solution should be light-touch, easy to use and compatible with their existing systems.
“That’s the key to real-world success.”
