Talkdesk’s general manager of healthcare and life sciences, Patty Hayward, discusses the use of ChatGPT, a generative AI tool, in healthcare contact centers. ChatGPT helps improve efficiency and effectiveness by supporting agents in tasks like summarizing calls and recommending dispositions. It also assists with managing complex contact center workflows. While acknowledging challenges and potential biases, Hayward emphasizes the benefits of personalized patient interactions and the importance of early adoption to enhance the patient experience and operational efficiency.
An executive from Talkdesk, a contact center solutions provider, discusses the utilization of generative AI in the healthcare industry and emphasizes the need for caution.
According to research by the National Academies of Medicine, before the COVID-19 pandemic, healthcare worker burnout had reached crisis levels. In 2021 alone, over 330,000 healthcare professionals in the United States left the industry, with many others considering leaving, as reported by Definitive Healthcare, a commercial intelligence company.
For healthcare organizations, burnout results in decreased productivity and increased recruitment and training costs. It also hinders the ability to deliver a superior patient experience when employees are overwhelmed and dissatisfied.
In today’s fiercely competitive market, where retail disruptors like CVS Health and Walmart Health are leveraging consumer-centric business models, subpar patient experiences can have detrimental effects on payers and providers, leading to reduced revenue and diminished market share.
Patty Hayward, the General Manager of Healthcare and Life Sciences at Talkdesk, a company specializing in healthcare customer-service technology and services, highlights generative AI as a valuable tool to address these challenges, particularly through ChatGPT.
Hayward explains that Talkdesk is currently using ChatGPT to enhance the efficiency and effectiveness of healthcare contact center agents. By combining AI with automation, they alleviate the burden on agents by handling tasks, creating a seamless experience for patients.
In an interview, Hayward discusses the application of ChatGPT in healthcare, how Talkdesk has integrated ChatGPT as an early adopter, the benefits of this generative AI tool, and the areas where healthcare organizations need to exercise caution.
Q: What makes ChatGPT particularly interesting in the health IT industry?
A: It is well-known that generative AI and ChatGPT face challenges in clinical settings today. Issues such as accuracy and “hallucinations” in large language models make the concept of a “ChatGPT Doctor” far from reality and potentially impossible.
In healthcare, AI works best when supporting human decision-making and training rather than replacing it. However, there are concerns related to patient privacy and the lack of governance that health IT leaders are currently navigating.
The contact center space, where we operate, provides an ideal environment to experiment with generative AI while delivering immediate value to healthcare organizations. Contact centers handle high volumes of often transactional interactions, such as checking claims, rescheduling appointments, or requesting prescription refills.
Healthcare has long sought to automate self-service in these areas, and generative AI allows us to provide more complex and relevant answers to patients, freeing up staff time for more intricate interactions. Healthcare conversations can be highly complex and require extensive knowledge and empathy from staff.
A negative contact center experience with a retailer may be frustrating, but in healthcare, it can impact health outcomes, and adherence to care plans, and even traumatize patients in critical moments. Generative AI can serve as a co-pilot for staff, offering answers and suggestions on tone and wording to support overwhelmed or inexperienced employees.
Enhancing the efficiency, accuracy, and empathy of these interactions benefits both patients and the organization.
Q: How has ChatGPT been integrated into your contact center workflow? How do agents’ roles change?
A: Generative AI excels when provided with specific inputs, such as transcripts or message drafts, using its language knowledge to generate different outputs. ChatGPT’s ability to take a concept and write a song or poem about it can be particularly delightful and entertaining.
In a healthcare contact center setting, numerous tasks require manipulating written information, such as summarizing, synthesizing, and personalizing it. Generative AI proves highly valuable in these areas.
Our initial focus was on tasks agents perform after calls or chats, such as submitting notes and entering dispositions for patients or members. Given that we have been supporting live transcription for calls and chats for years, we now offer agents a one-click summarization of any completed call. They instantly receive a standardized summary that they can edit if necessary or send directly to the system of record with another click. Our AI can also recommend a disposition based on the conversation, saving agents valuable after-call time.
We also employ AI to aid in managing and administering healthcare contact centers. Even in smaller provider or payer organizations, there can be considerable complexity related to facilities, departments, teams, and patient/member journeys. Generative AI dynamically suggests templates and support content for administrators, eliminating the need for extensive technical support documentation or direct contact with our team. We also leverage generative AI to propose chat responses based on context and knowledge-based articles.
In all cases, the overarching goal is to empower staff to become instant experts, whether during or after interactions with consumers or in contact center administration.
Q: How does ChatGPT contribute to your operations, and what are the benefits of using AI in this manner?
A: Generative AI surpasses rules-based configurations and static variables by providing personalized answers and information to consumers and staff in the most effective way. Currently, most self-service flows require constructing different templates based on specific data, such as patients’ names and appointment details. This process is time-consuming and often requires patients to adapt to the bot’s predetermined questions and input order. Generative AI, on the other hand, uses the same data to generate relevant and personalized responses immediately, reducing the need for managing numerous templates.
For patients, this means true natural language understanding and answers personalized to their unique information, health literacy level, and previous interactions with the healthcare organization. Patients desire a sense of familiarity and care throughout their journey, not just during face-to-face encounters with clinicians. Unfortunately, healthcare organizations frequently provide impersonal and formulaic experiences, forcing patients to repeatedly explain their needs and navigate the system. Generative AI can instantly translate information for staff or self-service, enabling it to be effectively utilized in conversations with patients, saving them time and reducing frustration.
Q: What precautions must be taken when using ChatGPT? Many experts have concerns about AI, such as its accuracy and biases.
A: As mentioned earlier, the current accuracy issues associated with ChatGPT raise valid concerns, especially in clinical settings. However, given the rapid evolution of this technology, it is expected that many of these accuracy issues will diminish shortly.
Nevertheless, ethical and moral considerations regarding the use of this technology are lagging behind its development. Are there inherent biases in ChatGPT? Does generative AI exacerbate data security and patient privacy concerns? It will be interesting to observe whether and when a regulatory framework emerges.
Generative AI algorithms have limitations, but they will continue to improve in accuracy. However, waiting for perfection is not a viable approach to implementing transformative technology. Healthcare organizations can leverage ChatGPT today to enhance the patient/member experience, improve operational efficiency, and reduce contact center agent burnout.
