AI Fails Without This

The future of home healthcare can’t rely on legacy systems and disconnected data.  It needs intelligent infrastructure that supports providers, not overwhelms them. 

In Episode 41 of #TeqTalk, Jas Kaur sits down with Venkatesh Guru Prasad, Director of Data Architecture & Governance at BAYADA Home Health Care, to discuss: 

✔️ Why nurses spend 50% of home visits documenting 

✔️ How BAYADA is modernizing its data platform with Databricks 

✔️ The risks of hallucinations in GenAI and how to avoid them 

✔️ What it takes to balance innovation, compliance, and empathy in care 

Key Takeaways: 

Fix the Foundation Before Scaling AI 

✔️ GenAI is not plug-and-play it requires fine-tuning, workflow integration, and oversight 

✔️ Most failures stem from missing infrastructure and siloed data 

✔️ BAYADA built scalable pipelines for real-time insights while protecting care quality 

“It’s not about throwing shiny AI tools. It’s about solving the right problem at the right time with the right data.” 

Document Less, Care More 

✔️ Nurses were spending more time on forms than on patients 

✔️ AI is now piloted to auto-generate visit summaries, chart notes, and decisions 

✔️ AI doesn't replace clinical judgment; it enhances. 

“We needed a system that supports clinicians, not adds to their burnout.” 

Listen to the Full TeqTalk Episode  

Whether you’re building in healthcare, leading IT, or implementing AI in regulated system this episode is packed with practical takeaways for real-world innovation. 

 Data Strategy > Data Dump 

✔️ Feeding 10+ years of data into GenAI won’t guarantee smarter output 

✔️ Relevance, context, and quality are what make models accurate and efficient 

✔️ Poor data selection leads to wasted compute and risky predictions 

“These models are smart. You just need to make them business-relevant.” 

Mitigating GenAI Hallucinations 

✔️ Strict validation pipelines and golden datasets are critical 

✔️ Human-in-the-loop sign-offs ensure safe, bias-free outcomes 

✔️ Salesforce, Databricks, and in-house platforms are monitored closely for drift 

“Confident AI isn’t always correct. That’s where governance matters.” 

What’s Ahead for BAYADA 

✔️ Scaling real-time decision intelligence 

✔️ Hands-free documentation across care teams 

✔️ Continuous EMR modernization to reduce tech burden 

✔️ Training clinical teams to trust and supervise AI 

“Innovation in healthcare can’t come at the cost of care.” 

See You Next Time 

At TeqTalk, we’re spotlighting voices building the future of healthcare one system, one insight, and one patient outcome at a time. 

Have a story or challenge to share? Connect with us on LinkedIn—we’re listening.