A GenAI Playbook for Healthcare Leaders

Healthcare organizations are rapidly moving forward with AI. The technology is opening up possibilities to improve operations, increase business efficiency and profitability, and drive better patient outcomes. 

According to Gartner, 79% of US healthcare payers indicate that they have started, or will start within the next two years, formal use-case assessments and copilot solutions from enterprise vendors to bring GenAI capabilities into the organization. Another survey shows a striking majority of healthcare organizations (78%) are formally assessing use cases, while 66% are developing the ethics and policies for appropriate use and 47% are funding initiatives. 

If you’re not planning on implementing AI, you’re already falling behind. This playbook for healthcare leaders will help you know what to ask your CIO/CTO, which criteria to use to decide on use cases, and how to prepare for success with GenAI.  

While AI brings many advantages, some healthcare leaders are hesitant to implement the technology due to some of the technical challenges. Accuracy and reliability can be difficult; you’ll have to stop hallucinations and make sure your AI is grabbing data from the right places. If you’ll be deploying AI widely, it needs to be done carefully so it doesn’t run up high expenses. Additionally, data privacy—both for your organization and for patients—has to be a priority. 

You know you’re ready to implement GenAI when you’re ready to address those challenges. We wrote about how to understand and avoid those pitfalls in this blog post. If you solidly understand how to navigate those hurdles and you’re ready to build, here’s what to do next. 

While GenAI is a useful tool for accelerating business, it’s not a comprehensive solution that will magically solve all your company’s problems. Its greatest impact comes from combining it with other technologies and user interfaces. To make sure you’re getting the most value out of your GenAI, be sure to prepare your organization by identifying use cases ahead of time. GenAI is well suited for use cases such as: 

• Content generation. GenAI has a particular strength in content generation. Use GenAI to create personalized member communications and to take notes and summaries of previous member interactions. You can also use it as an ambient digital scribe for call service reps and care management specialists.  
• Discovery. Healthcare leaders can use GenAI to help members find care and services in their area based on their unique needs. Members can compare different plan options to find the plan that’s best for them. Member onboarding and services enrollment is also an area where GenAI is beneficial as it helps members discover your offerings. If you need to consult medical processes, understand eligibility data, and see if prior authorization is required, GenAI can help.  
• Decision making. If healthcare leaders or patients have questions, decision-making GenAI can provide answers. Healthcare companies can use GenAI to schedule appointments on behalf of members and see available appointment times, create member-friendly EOBs and provider bill audits, refill prescriptions, transfer data between healthcare providers, and provide information to help members make decisions based on their unique situations.  
• Data extraction + summary. Use GenAI for member chatbots to answer member questions, explain plan details and ancillary benefits, give you insights into plan usage, and perform data analysis and reporting on plan and member insights. 

Once you’ve identified possible use cases for your organization, you can assess the use cases to see which ones would be most easily implemented and which ones would provide the greatest value. Healthcare leaders can evaluate each proposed use case against the following criteria to select the most impactful use cases. 

One major consideration for use cases is technical feasibility. The use case needs to be compatible with your organization’s existing data architecture and systems. You’ll also need to understand how implementing this use case of GenAI would impact your existing operations and how products and product features could potentially be affected. 

As you’re doing your due diligence on potential use cases, you’ll also want to consider speed to value. How quickly can you stand up this particular use case? How much labor and money—both up-front and over time—would it take for this use case to start providing value? Similar to speed and value, you’ll also want to consider return on investment. Each potential use case should have an estimated ROI and an approximate timeline of when the use case would start realizing that ROI. 

Your assessment criteria should also include an analysis of how this use case would fit into a broader AI strategy. Certain use cases of GenAI may be a good idea but may not fit with the organization’s broader AI and corporate strategy. Healthcare companies need to be especially aware of data privacy and other risk factors that may need to be addressed as part of their AI strategy. 

You’ll need to evaluate the technical approach to your use case carefully, because it could impact the cost, accuracy, and effectiveness of your project. Work with your CTO to consider some of the following technical components: 

• Language Processing Units (LPUs)
  • How are we selecting the LPUs with the right capabilities while controlling run-time costs?
  • What strategies will we employ to handle complex processing tasks such as planning, reasoning, and agentic behavior?
• Orchestration
  • How will we integrate and manage the workflow for various components within the AI application system?
  • How do we build the AI application so you can swap in different components to evolve AI capabilities?
• Memory
  • How do we build conversational memory across multiple member interactions and other domain-specific knowledge bases so the AI applications become increasingly personalized and useful?
• Infrastructure
  • What deployment infrastructure is needed to ensure scalable and reliable application management?
  • How can observability tools be leveraged to monitor and maintain the performance of our AI system?
• Data Engineering
  • Is our data ready to support the AI model and consumption? What additional data acquisition, integration, and transformation is needed?
  • How do we design robust data pipelines and fine-tuning mechanisms to support the AI system?
• Services + Tools
  • What enterprise systems and productivity tools need to be integrated with the AI application to create a bi-directional feedback loop?
  • What project management and optimization services do we need to ensure comprehensive support and enhancement of the AI platform?
• Security + Compliance
  • What security measures are necessary to protect sensitive data processed by the AI system?
  • How will we ensure the AI platform complies with relevant regulatory standards and legal requirements? 

How do you know how much up-front investment it would take to implement GenAI? What GenAI options are compatible with your existing data architecture? As you consider questions and use cases related to implementing GenAI, 3Pillar can help. To get tailored support as you find the most profitable use cases and identify the projects with the fastest speed to value, contact us today.