Leveraging Generative AI to Enhance Patient Scheduling Experience

Key Takeaways:

1. Prioritizing Artificial Intelligence (AI) is crucial for small and medium-sized businesses (SMBs) in 2025, with 35% of SMBs listing AI as a top technology investment priority.
2. MD Fit, a healthcare technology company, has successfully integrated generative AI into their patient scheduling platform to improve the patient experience and operational efficiency.
3. MD Fit's key challenges included transcription errors, complex scheduling tasks, and the need to provide a seamless experience for patients who prefer to interact with a live agent.
4. By integrating Amazon Lex and Bedrock, MD Fit was able to address transcription errors, enhance scheduling workflows, and maintain traceability for regulatory compliance.
5. The initial implementation of the automated solution resulted in a 60% reduction in the need for live agents, leading to immediate return on investment for their launch customer.
6. Future plans include implementing more natural conversation flows for complex scheduling tasks, enhancing appointment rescheduling, and ensuring appropriate guardrails for medical advice.
7. The key to success is to identify a business problem, experiment with technology, and iterate on the solution in a cost-effective and scalable way, with the support of the AWS account team and partners.

Prioritizing AI for SMBs

• According to IDC's worldwide SMB survey, 35% of SMBs listed AI as a top technology investment priority for 2024, more than double the 16% seen in 2023.
• SMBs are taking a "fast follow" approach, learning from earlier adopters before investing in AI technologies themselves.
• By 2025, 70% of SMBs will demand clear use cases before investing in generative AI (Gen) or AI technologies.

MD Fit's Journey with Generative AI

About MD Fit

• MD Fit brings together patient and provider information into a single system to ensure the right patient is seen by the right provider at the right time.
• The platform consists of multiple applications, including Advisor (call center scheduling), Patient (web-based self-service), and Automate (voice-based self-service).

Challenges and Solutions

• Challenges:
  • Transcription errors in the Automate voice-based self-service solution
  • Complex scheduling tasks, such as new patient registration and appointment rescheduling
  • Providing a seamless experience for patients who prefer to interact with a live agent
• Solutions:
  • Integrated Amazon Lex and Bedrock to address transcription errors and enhance scheduling workflows
  • Maintained traceability by keeping both original and corrected outputs
  • Allowed for quick iteration, with a single junior developer implementing the Bedrock-powered workflows in two sprints

Results and Future Plans

• Initial implementation of the Automate solution resulted in a 60% reduction in the need for live agents, leading to immediate ROI for the launch customer.
• Future plans include:
  • Implementing more natural conversation flows for complex scheduling tasks
  • Enhancing appointment rescheduling
  • Ensuring appropriate guardrails for medical advice

Getting Started with Generative AI

• Identify a business problem, experiment with technology, and iterate on the solution in a cost-effective and scalable way.
• Engage with the AWS account team and partners to access resources and accelerate the implementation.
• Align generative AI ideas with business priorities and build an experimentation culture.

Key Advice from Sean O'Brien

• Focus on solving a business problem with technology, not fitting a technology into the business.
• Start with a specific, high-impact area and iterate, rather than a full platform rewrite.
• Engage the AWS account team to access resources and partner support to accelerate the implementation.