🎙 Milind Sawant: AI Deployment in the Healthcare Sector

My guest today is Milind Sawant, a leader in product innovation with 30 years of global R&D experience.

At Siemens Healthineers, he founded the Center of Excellence for AI and Machine Learning, growing it to more than 200 cross-functional team members and achieving significant business impact.

Milind shares a structured approach to AI deployment, covering common pitfalls, a problem-first adoption framework, and strategies for balancing speed and accuracy.

We also discuss the evolving role of subject matter experts, overcoming resistance to change, and the future role of AI in healthcare.

His insights are practical, experience-driven, and invaluable for anyone looking to adopt AI in their business.

I hope you enjoy listening to our conversation.

- Sam

🎙 PODCAST EPISODE #4

Listen on your podcast platform of choice:

✍ MY TAKEAWAYS

How can companies avoid common pitfalls when deploying AI?

• Start with a problem-first approach to avoid data accumulation without purpose.

• Involve subject matter experts (SMEs) alongside IT teams to ensure contextual relevance and effective problem-solving.

• Prioritize data quality over quantity, ensuring labeling accuracy to prevent misleading AI predictions.

• Avoid unnecessary model complexity that increases computational demands without corresponding benefits.

• Maintain a strong in-house AI capability to avoid excessive reliance on external vendors who may lack domain expertise.

Why is a problem-first approach essential in AI projects?

• Identifying a clear problem statement ensures AI efforts directly address business needs and justify resource allocation.

• This approach guides data selection and model complexity, aligning AI initiatives with specific business outcomes.

How can subject matter experts drive AI success in organizations?

• Engage SMEs from relevant departments to guide problem definition, data preparation, and model interpretation.

• Form cross-functional teams with SMEs and AI data scientists to align AI solutions with domain-specific knowledge.

What role does data quality play in the success of AI projects?

• High-quality, accurately labeled data is critical to training effective AI models and achieving reliable outputs.

• Poor data quality leads to incorrect predictions, especially in critical applications like healthcare.

How can pilot projects generate momentum for AI adoption?

• Select projects that are small, quick to complete, and use accessible data to demonstrate value and build excitement.

• Focus on achieving quick wins to maintain enthusiasm and justify further investment in AI initiatives.

How do businesses balance the trade-off between speed and accuracy in AI implementations?

• Begin with simpler projects to minimize risk and enable quick iterations as understanding and capability grow.

• Adopt an iterative approach, refining data and models at each step to gradually improve accuracy.

How can organizations overcome employee resistance to AI?

• Emphasize AI as a tool to enhance, not replace, existing roles, and encourage upskilling for future relevance.

• Provide training to help employees leverage their domain expertise in conjunction with AI technologies.

What AI training tools are available for mid-career professionals?

• Focus on understanding AI concepts and their application to your domain, rather than programming skills.

• Use no-code tools like Orange Software or explore platforms that integrate AI functionalities without requiring deep technical expertise.

How should business leaders strategize AI investments?

• Direct AI efforts towards projects that improve end-customer experiences to create a positive business impact.

• Ensure projects align with real customer needs to enhance product offerings and drive revenue growth.

Where might AI transform patient care and clinician roles over the next decade?

• AI initially will assist clinicians by managing routine cases, allowing healthcare professionals to focus on complex scenarios.

• As AI matures, it may handle more standard cases independently, with clinicians overseeing edge cases and validation.​

👋 THAT’S ALL, FOLKS

FEEDBACK

If you’ve made it this far, I’d love to know what you thought of this edition—or others. Good or bad, give me your feedback below! 👇

What do you think of Not a Threat?