VB Group Interview Transcript | How iFLYTEK Strategizes in Smart Healthcare and Its Product Portfolio

In 2017, the hottest topic was undoubtedly artificial intelligence (AI). As China’s largest company specializing in intelligent voice and AI technologies, Iflytek has drawn significant attention for its every move. Beyond its well-known intelligent voice products, Iflytek had already begun to strategically position itself in the smart healthcare sector. Since 2015, leveraging its strong AI R&D team and technological advantages, Iflytek has not only applied intelligent voice technology to the medical field but also expanded its business to include medical imaging and cognitive computing-based clinical decision support systems.

Recently, we invited Dr. Tao Xiaodong, CEO of iFlytek Medical, to participate in an interview with the VB Group.

Founded in 1999, precisely during the second AI winter, iFlytek has been dedicated to developing the latest and most cutting-edge technologies in the field of artificial intelligence. Despite encountering numerous challenges, the company has persevered to the present era of widespread AI adoption. iFlytek has amassed substantial expertise in core AI technologies, enabling it to integrate these capabilities with various industries to deliver practical applications that effectively address real-world problems.

At its inception, Iflytek primarily focused on speech-related technologies. Throughout this process, it accumulated extensive experience in core artificial intelligence (AI) technologies and general-purpose algorithms, laying a solid technical foundation for today’s integration of “AI + industries.” From 2006 to 2017, Iflytek’s speech synthesis technology ranked first globally for 12 consecutive years, and it was the first to surpass human performance in terms of naturalness in synthesis technology.

In the field of speech recognition, iFlytek ranked first globally in accuracy across three different scenarios (single-microphone, dual-microphone, and six-microphone array) at the 2016 CHiME-4 challenge. Furthermore, iFlytek secured the top global positions in two other significant competitions: commonsense reasoning and knowledge graph construction. In June of this year, iFlytek was ranked first in China and sixth worldwide on the list of the World’s 50 Smartest Companies. These achievements demonstrate that iFlytek’s expertise and accomplishments in the field of artificial intelligence have gained widespread recognition within the industry.

Even so, Tao Xiaodong stated, “We still have a long way to go.”

At the end of 2015, iFlytek began to lay out its presence in the healthcare industry and established a dedicated subsidiary, iFlytek Medical, in 2016.

“iFlytek Healthcare’s primary mission is to leverage artificial intelligence technology to support China’s healthcare reform and contribute to the building of a Healthy China,” stated Dr. Tao in the interview. “In this context, we have identified certain pain points in China’s healthcare reform and current medical landscape, such as the scarcity of high-quality medical resources and the uneven distribution of medical expertise. How can these issues be addressed? It is difficult to resolve them quickly through conventional means, including internet-based approaches. Therefore, we aim to harness AI technology as a tool to assist physicians, enabling those performing at an average level (‘60-point doctors’) to achieve a significantly higher standard (‘80-point doctors’).”

iFlytek primarily focuses on several key initiatives in the healthcare sector. The first is leveraging iFlytek’s language technologies to enhance physicians’ efficiency in documenting medical records. Physicians typically spend approximately 40% of their time writing medical records, and the conventional mouse-and-keyboard input method is often cumbersome. Amidst their demanding daily schedules, clinicians require more efficient tools to complete this administrative workload. To address this, iFlytek has developed a voice-enabled electronic medical record (EMR) system based on speech interaction technology. By integrating automatic speech recognition and semantic understanding, the system directly converts physicians’ dictated content into structured medical records. Currently, the most mature applications are in dentistry and ultrasonography. Dedicated microphones have also been designed to allow physicians to wear them for continuous audio capture throughout consultations. This voice-based EMR entry significantly improves documentation efficiency, saves substantial time, and enables physicians to devote more attention to patient care.

Triage nurses in hospitals must answer a large volume of patient inquiries every day, many of which are simple and repetitive. For instance, at a relatively busy Grade 3A hospital, a triage nurse may need to respond to nearly a thousand questions from patients or their families daily. The vast majority of these inquiries concern hospital locations and procedural guidance. As these questions are straightforward and repetitive, they can be fully handled by machines, especially when supported by advanced voice interaction technologies. To address this, iFlytek has developed a triage robot. After being deployed at the PLA General Hospital (301 Hospital), the robot engages in approximately 500–800 interactions per day. This allows triage nurses to be freed from handling simple, repetitive queries and instead focus on more professional tasks, such as providing medical guidance and determining the appropriate department for patients based on their described symptoms.

By integrating the hospital’s inpatient information systems, a comprehensive view of patient data can be consolidated into a single mobile client application. This enables attending physicians to access real-time, complete patient information—including laboratory and diagnostic test results, medication administration, and nursing care records—facilitating timely adjustments to medical orders and accurate documentation of the patient’s clinical course. Furthermore, these operations can be executed via voice commands, supported by intelligent speech recognition technology. Leveraging this complete and accurate panoramic patient data, the system can also perform intelligent analytics to push relevant clinical guidelines, latest literature, and recent research findings to physicians, thereby providing evidence-based recommendations for diagnosis and treatment.

In the medical field, some of the most prominent applications of artificial intelligence (AI) are in medical imaging. In fact, AI has been developing in the area of computer vision for many years, and the concept of Computer-Aided Diagnosis (CAD) was proposed decades ago. However, with the recent surge in popularity of AI—including deep learning algorithms and the vigorous development of imaging-assisted diagnosis—iFlytek has also made strategic moves in medical imaging. Like most research institutions and startups, iFlytek started with pulmonary nodules. On one hand, there is a substantial amount of public data available to support this effort; on the other hand, lung cancer has a relatively high incidence rate in China. Moreover, iFlytek’s average detection rate has reached 94.1%. In addition, iFlytek has developed solutions for mammography and brain MRI.

Dr. Tao stated, “There is ongoing debate regarding the actual accuracy of lesion detection and its practical applicability. For instance, an accuracy rate of 90% may offer utility commensurate with that level, while a 95% accuracy rate may provide corresponding benefits. The key lies in identifying appropriate clinical scenarios to determine which solution can truly assist physicians.”

At the end of the interview, Dr. Tao also remarked, “The greatest beneficiaries of AI’s development in the healthcare sector are primary care physicians, as it can assist them with image interpretation, diagnosis, and treatment. However, there is also a challenge: primary care physicians may not have a clear understanding of the most cutting-edge technologies and may be unaware of which pain points can be addressed by AI.”

Thus, it appears that the key to developing AI technology in the healthcare sector lies in identifying appropriate directions and integrating them with practical application scenarios.

The above content is a transcript of the VB Group interview live broadcast. To listen to the full live broadcast, please enter the live room.：https://m.qlchat.com/topic/details?topicId=2000000005714742