AI-based clinical assistant Kahun raises $8m

• Author: Gali Weinreb
• Published date: 02 October 2022
• Publication title: Globes (Rishon LeZion, Israel)

Current AI tools for healthcare providers have failed to address current challenges and gain trust in the medical community. They use a combination of big-data engines built on patient records and experts' knowledge, not on evidence-based medical literature like Kahun's tool

The company was founded by a team that mostly comes from the technological world. CEO Eitan Ron and CTO Tal Goldberg founded Human Click, which was acquired by LivePerson (Nasdaq: LPSN; TASE: LPSN) for $9 million, which then represented 15% of LivePerson, which today has a market cap of about $700 million. Ron and Goldberg became executives at Liveperson, and Goldberg even went on to various positions at Waze, now part of Google. But a few years ago the two decided to open their own company again together with chief medical officer Michal Tzuchman Katz.

30 million connections that model the medical data available in literature

In contrast to some of the other companies in the field whose algorithm is built by scanning huge databases of medical records, Kahun's algorithm is based on the professional literature in the medical sector. Based on a review of medical articles, the company has built a map containing 30 million connections that, it believes, represent the medical knowledge available in the literature, and its algorithm runs on this map.

Ron said, "This is a significant distinguishing element of our company's technology. When we propose a diagnosis or treatment method, we can immediately refer to the source in the literature where this idea appears, and the doctor can examine the quality of the literature that offers the same solution.

"This, compared with a system that runs on medical records and derives insights from them, but it acts as a 'black box', and we humans do not know how to criticize its recommendations or neutralize its biases."

What deficiency is this system designed to fill? Lack of medical personnel? Or providing medicine in a place where there are no doctors?

"No, we don't replace a doctor, but help them. In the first step, we help doctors think. Even today, they use a computer to think, but they conduct a search that only bears fruit if they already have some kind of hypothesis about what is going on."

A mass application was not profitable

The company first followed the accepted business models of the content application field. The product was released to the market in early 2021 and was initially offered free, to encourage use. The app reached about 7,000...