Healthcare organizations are facing immense pressure to do more with less. And with the explosive developments in content-generative AI, marketing, digital strategy, and patient experience teams are looking to ChatGPT and other large language models as a method to relieve some of this pressure.

The healthcare industry can benefit from generative AI technology by leaning into its power of streamlining digital content creation. At the same time, healthcare organizations should be aware of when to utilize these transformative technologies given the sensitive nature of cultural nuances and personal health data they work with every day.

When should generative AI be used in healthcare?

Healthcare content can be categorized into two categories:

• Informational: Evergreen content that answers common questions or provides solutions users seek to fulfill a goal or overcome a challenge.
• Branded: Content designed to evoke feelings, inspire thoughts, and generate consumer awareness by associating a brand with its value proposition. Next, we identify when generative AI can empower healthcare digital content teams, and where there still are some challenges with this quickly evolving technology:

Informational content

As of today, the informational content type poses the most opportunities for generative AI to efficiently streamline content outputs. Because informational content is considered more prescriptive than branded content, AI won’t require the same learning curve to generate drafts as it would more nuanced information.

Examples of healthcare content that generative AI can produce, include:

• Services (how do healthcare providers diagnose and treat heart conditions?)
• Conditions (what are common heart attack symptoms?)
• Procedures (what is a trans-aortic valve replacement?)
• Testing (how do I prepare for an echocardiogram?) However, with the potential of every healthcare organization integrating AI into their content generation process, how do we lean into creating differentiation? The best way to accomplish this is through a robust branding strategy enriched with tone and voice guidelines, as well as a strong editorial style guide.

Generated content will also still require human review for accuracy and authenticity and to review for nuanced content, for example, scanning for anything that might have implications for equity or inclusion.

Content AI can also provide initial drafts of content that already exists within the organization’s digital ecosystem. What’s important to note is that “initial drafts” mean creating a foundation or jumping-off point for content creators to further differentiate and refine. This potentially could include:

• Blog content (thought leadership pieces)
• Awards and recognitions, rankings
• Insurance information and FAQs
• Billing information and FAQs
• Career sites: Job descriptions
• Page titles and descriptions
• Conditions, treatment, and services
• Physician bios 

Branded content

Simply put, branded content is unique to the brand itself. Strong brands produce messaging that drives engagement by speaking directly to an audience segment’s goals, challenges, and behaviors. Tone and voice differentiation is created by:

• First, building a solid voice foundation that anchors the brand. How does the voice carry the brand? The voice is constant.
• Second, creating tones that scale per audience segment and meet each segment where its people are across various marketing channels. The tone is variable. 

In order to be successful, generative AI must learn about your brand through a robust set of brand guidelines. That’s why, now more than ever, healthcare organizations need to invest in tone, voice, and messaging guides that can support this. Should healthcare organizations turn to content AI for branded content, it’s crucial that all content is reviewed by a human content professional first before publishing. This is important because:

• Like a brand’s tone, content circles around context. Humans can deem when it is and isn’t appropriate to leverage generative AI for content outputs.
• AI must learn a brand framework and how it relates to a target audience. Simply publishing informational content without a branded angle that speaks to an audience’s goals, challenges and behaviors risks healthcare org differentiation in a dense market.
• AI hasn’t fully grasped syntax yet (how grammar is arranged based on a writer’s style and known grammar rules). 

When shouldn’t generative AI be used in healthcare?

First and foremost, AI should never be used in any scenario that would be in direct violation of HIPAA, including protected information like PII/ PHI.

Areas that we propose caution before considering generative AI is when the human touch is still essential, such as understanding cultural nuances or ensuring content isn’t considered offensive or performative.

Examples of healthcare content that generative AI should not create:

• Patient and donor impact stories
• News stories, press releases, or crisis communications (see Vanderbilt example)
• Legal disclaimers, and policies

Whether healthcare organizations are leveraging AI to accelerate content production or to more consistently express their brand, understanding when to appropriately use it and how will be key to a successful implementation.

We hope you enjoyed reading the Phase2 blog! Please subscribe below for regular updates and industry insights.