From Data to Trust: Preparing Healthcare for AI Adoption

AI has the potential to transform healthcare — but only if it is built on strong foundations of data quality, safety, and trust. In a recent vodcast interview, Victor Lazarevich, CTO of Digiteum, shared his perspective on what it takes for healthcare institutions to adopt AI responsibly and effectively.

With more than 20 years of experience in digital transformation, Victor has guided organizations such as Diaceutics and others in designing scalable data and AI solutions that improve daily operations and deliver lasting impact.

Watch the full interview below or explore the key takeaways from the conversation.

“The most uplifting moment is when people actually start using the software — and asking for more.”

Victor explained that building advanced solutions means little unless they truly change how people work. The moment clinicians start requesting new features is when technology proves it’s delivering value.

In healthcare, mistakes can cost lives. That’s why every AI system must:

• Go through clinical trials and peer reviews.
• Adhere to strict national and industry standards.
• Distribute responsibility across vendors, institutions, and clinicians.

Trust isn’t optional — it’s the foundation.

GDPR and Data Protection as Enablers

Many see GDPR as a blocker. Victor calls it a stepping stone. Key principles include:

• Anonymization: removing identifiers like names or addresses.
• Data minimization: using only what’s necessary.
• Relevance: feeding models the right data for the right decision.

Preventing the “Black Box”

To win trust, AI must be transparent:

• Share confidence levels (e.g., 70% likelihood of diagnosis).
• Provide simple reasoning.
• Position AI as an assistant, not an unquestionable authority.

Integration Is Everything

If staff must juggle multiple systems, adoption will fail. AI needs to:

• Integrate with existing EHR systems.
• Provide a seamless user experience.
• Include feedback loops so users feel ownership.

Driving Adoption

Resistance comes from skepticism, complexity, or bias. The solution?

• Training focused on real use cases.
• Finding AI champions inside organizations who lead by example.

Data Quality: The Non-Negotiable

No good AI without good data. Steps include:

• Regular data audits.
• Breaking down data silos.
• Establishing governance practices.

Start Small, Scale with Proof

Forget the “perfect” business case. Begin with a pilot project, measure impact, and then expand.

Keeping Models Relevant

AI models degrade without updates. Hospitals need:

• Teams monitoring performance and accuracy.
• Standards for collecting fresh, relevant data.
• Collaboration with vendors to improve continuously.

Healthy Collaboration with Vendors

Strong partnerships require:

• Clear contracts on IP, ownership, and responsibilities.
• Open standards to avoid vendor lock-in.
• Step-by-step growth, based on results and trust.