This article is sponsored by Nuance, a Microsoft company. Discover clinical documentation that writes itself™.

There’s no doubt about the biggest story in tech over the last year: the rapid rise of generative AI. Although the underlying technology that powers generative AI has been around for many years in various forms, when OpenAI’s ChatGPT was made publicly available, it launched a wave of headlines around the world. And just like any technological breakthrough, a wave of hype soon followed, with hundreds of organizations claiming their generative AI app would change the world.

Many of these organizations focused on the health care sector, aiming to help health care facilities use generative AI to overcome key challenges. But savvy clinicians know it’s vital to assess a vendor’s true capabilities rather than get swept away by the hype and put care quality, clinician wellbeing, and patient safety at risk.

Artificial intelligence: a quick primer 

Before we look at the factors that health care leaders should consider when assessing generative AI solutions, let’s take a moment to clarify some of the terms used (and sometimes misused) in the AI world:

• Artificial intelligence: An umbrella term describing a machine’s ability to perform tasks that would typically require human intelligence.
• Machine learning: Algorithms trained to detect patterns in large volumes of data to suggest actions and predict outcomes.
• Deep learning: Machines that mimic the operations of the human brain to process multiple data types and learn faster with less direct intervention from trainers.
• Conversational AI: Systems that understand the meaning, intent, and sentiment of users’ natural language and offer relevant, conversational responses.
• Ambient AI: Machines that monitor their environment to provide intelligent assistance to users when needed, without having to be specifically prompted.
• Generative AI: Deep learning models pre-trained on vast amounts of data, enabling them to produce new content in response to user prompts.

Over the last couple of decades, advances in machine learning and deep learning have had a significant impact on health care.

AI has transformed medical imaging, augmenting radiologists’ expertise by spotting diagnostic clues that the human eye might miss. Machine learning algorithms have revolutionized drug discovery by rapidly identifying patterns in enormous quantities of data that would take human researchers years to uncover. And some health systems have begun using AI-powered tools to predict likely spikes in demand, helping managers allocate scarce resources where they’re most needed.

Conversational AI applications that allow clinicians to dictate notes directly into the EHR have dramatically reduced the documentation burden. More recently, ambient AI solutions have emerged that can capture the full patient story at the point of care without the need for dictation. Some of the most advanced ambient AI tools can even help improve care delivery by identifying social determinants of health, for example by analyzing patients’ speech for biomarkers indicating depression or anxiety.

And now, generative AI is adding to these conversational and ambient AI capabilities by enabling systems to automatically draft clinical notes in seconds and make them available for physician review immediately after each appointment.

As AI technology in all its forms continues its rapid evolution, it will have a profound impact on every aspect of health care—from rare condition research and early disease detection to clinical decision support and personalized medicine.

But this future of AI-augmented clinicians delivering high-quality care and better patient outcomes is only possible if vendors have the right combination of technology, expertise, experience, and scale.

Five things to look for in a health care AI technology partner 

1. AI fine-tuned for health care workflows. Widely available generative AI models can provide raw power to analyze data and generate responses. But unless the applications built on these models are tailored for complex, interconnected health care workflows, they’ll struggle to deliver meaningful value. Look for vendors with a record of delivering trusted technology solutions that are relied on by clinicians and support staff in their everyday workflows.

2. A responsible approach to AI. Perhaps more than any other industry, AI in health care must be built responsibly and ethically. Good health care relies on using highly sensitive patient data and making decisions based on clinical evidence. So, vendors should have a strong ethical AI framework that ensures products are built—and used—responsibly. Ask potential vendors to share the details of their ethical AI framework with you to make sure they’re taking their responsibilities seriously.

3. Deployment and optimization expertise. Many generative AI startups have great ideas. But actually, deploying applications in the real world, and tightly integrating them with EHRs is a very different ball game. Then there’s the question of whether a startup will have the ability (or the longevity) to support customers to continually optimize their deployments to deliver maximum long-term value. Check that your prospective vendor can share examples of large-scale technology deployments. And ask if they provide customer success managers or other support to help you maximize the value of your investment.

4. Enterprise-grade dependability. To provide the security, stability, and scalability that health care leaders need, AI vendors must have a trusted global infrastructure that has ultra-reliability, ironclad cybersecurity, and strong data governance at its heart. The best vendors will be able to demonstrate their security and governance credentials, and provide verifiable details about availability SLAs.

5. Deep health care experience. The best health care AI vendors will have a long track record of working in the industry and deep partnerships across the health care ecosystem—from EHR and academic research institutes, to health systems of all sizes. Look for vendors that can show they’re deeply embedded in the sector and understand the challenges and priorities of clinicians.

The AI-powered future of health care 

As the application of AI expands across the health care industry, some of the earliest gains have come in the way of faster, easier administrative tasks and clerical work. In an age where physicians are being asked to do too much in too little time, these solutions meet an immediate and growing need.

Nuance and Microsoft recognize the importance of using AI to automate tasks that impede care delivery. Technology is at its best when it solves problems, taking over tasks and allowing humans to perform better. Discover clinical documentation that writes itself™.

By working with trusted technology partners using solutions proven in real health care workflows, organizations can harness the very best of what AI advances have to offer to accelerate advancements in health care.

Rebecca Schechter is senior vice president and general manager, Dragon Ambient eXperience (DAX) and oversees Nuance’s DAX growth strategy, partner and customer relationships, and newly-centralized DAX operations. Rebecca has expertise in driving large-scale growth and operational strategies for rapid global expansion, accelerating innovation, and building strong customer and partner relationships.

Prior to joining Nuance, Rebecca served as the CEO of Optum Behavioral, as well as the executive vice president of benefits at Liberty Mutual. She also gained deep global experience during her time at McKinsey & Company, Thomson Reuters, and State Street where she worked and lived across Europe, Asia, and North America. She holds a bachelor of commerce in international business from McGill University and an MBA from Massachusetts Institute of Technology. Rebecca lives outside Boston with her husband and two children.