Google and Clalit develop an AI for 5 million patients
Google.org, Google's philanthropic arm, and Clalit Health Services have announced a two-year grant that will enable Israel's largest healthcare provider to develop and launch an AI-powered “digital front door”: an intelligent conversational assistant designed to make healthcare services more accessible, particularly for patients from diverse communities.
The system will function as an intelligent digital interface, guiding patients throughout their healthcare journey and providing continuous, personalised support. Built on Google Cloud's secure infrastructure and powered by advanced artificial intelligence, this comprehensive multilingual assistant will help patients better understand their healthcare needs, prepare for medical appointments and manage follow-up care in their native languages.
Through voice-enabled tools, Clalit will be able to communicate with patients using devices they already own, ensuring that access to healthcare is no longer hindered by limited internet connectivity or the absence of a smartphone.
“Our commitment to every individual, regardless of their background or community, is at the heart of everything we do. The support of Google.org and Google Cloud enables us to harness some of the world's most advanced technologies to ensure that every patient receives the best possible care, in their own language and tailored to their needs. The model we are developing here in Israel is intended to become an open-source architecture, accessible to healthcare organisations worldwide, allowing them to replicate it and contribute to ensuring access to healthcare for all,” said Eytan Wirtheim, CEO of Clalit Health Services.
Recognising Israel's linguistic and dialectal diversity, Clalit is committed to ensuring equitable access to healthcare for all patients. Traditional digital platforms, which rely primarily on formal written language, can exclude people who communicate mainly through local dialects.
As a result, online appointment booking rates remain low. Using large language models (LLMs) and Google Gemini, the new system will offer voice recognition adapted to local dialects and cultural nuances, enabling patients to easily complete registration, triage and appointment-booking procedures in their native languages.
The system will also address accessibility barriers faced by patients without an internet connection, as well as those using devices that restrict access to web applications and text messaging services.
By integrating an interactive, voice-controlled AI system directly into landlines and standard mobile phones, Clalit aims to eliminate these communication barriers. These personalised communication strategies are designed to increase enrolment in digital services and improve appointment confirmation rates across a wider range of communities.
To protect patient privacy, Clalit applies a Zero Trust security framework and stringent data protection measures. All interactions between patients and the conversational AI system are encrypted, stripped of personally identifiable information and fully anonymised.
Following an initial pilot phase, the system will be rolled out nationwide. Ultimately, Clalit plans to make its voice AI architecture and cultural adaptation frameworks publicly available as an open-source model for public healthcare, enabling healthcare systems around the world to replicate the approach.
Caroline Haïat
Comments