Artificial Intelligence does a lot of incredible things in healthcare – but it is still a newer tool that we need to be careful with. Without proper preparation, guidelines, and oversight, AI can run on bad data, create hallucinations and bias, leave your organization open to security threats, and much more. One of the very first steps you need to take to prepare your organization for AI solutions is proper data management. AI needs an incredible amount of data in order to function, and feeding it any data that is bad, improperly coded, etc., can cause a chain reaction of consequences, ranging from unfortunate to catastrophic.
To help you know how to take this first step towards AI solutions, we reached out to our talented Healthcare IT Today Community to ask — what are some of the key data management efforts healthcare organizations should be doing to ensure they’re ready for AI solutions? Below are their answers.
Elevsis Delgadillo, SVP, Customer Success at KeenStack
Healthcare organizations should invest in enhanced data governance immediately, if they are not already. Strong data governance is critical to the success when using AI solutions. AI outputs are only as good as the data you feed them, so poor-quality inputs will lead to poor-quality outcomes.
Sujay Jadhav, Chief Executive Officer at Verana Health
AI is only as effective as the data it uses. To prepare for AI solutions, healthcare organizations must focus on clinically guided curation that transforms raw, unstructured information such as physician notes and pathology reports into structured, interoperable formats. This also means linking EHR data with claims, imaging, or genomics to provide a more complete view of the patient journey. The true value is not in collecting more data, but in ensuring it is curated, clinically contextualized, and continuously validated to support decisions that improve patient outcomes.
Justin Miller, Director of Software Engineering at Full Spectrum
One of the most critical initiatives for AI-readiness is developing a strong data governance framework to ensure data is used ethically, securely, and in compliance with HIPAA and other regulations. This, of course, includes data access controls and audit logging, but also should include less obvious elements such as data anonymization, bias detection, and mitigation. Model transparency is especially important for clinical applications to build trust and facilitate adoption.
Basia Coulter, Partner, Healthcare and Life Sciences at Globant
AI readiness starts with robust data governance. This includes clear policies for data ownership, access, and quality assurance. Organizations should prioritize data standardization using common ontologies (e.g., SNOMED CT, LOINC) and invest in metadata management to track data provenance and usability. Building centralized data repositories or federated architectures with controlled access can facilitate model training without compromising privacy. Increasingly, this also means investing in the creation of data products, well-defined, reusable datasets curated for specific analytical or clinical purposes. Finally, routine auditing and validation of datasets ensure that AI models are trained on reliable, bias-aware data, which is essential for ethical and clinically relevant outcomes.
B.J. Boyle, Chief Product Officer at MacroHealth
To be ready for advanced analytics and AI solutions, healthcare organizations should focus on dismantling data silos and improving data liquidity. Information across the healthcare industry is often fragmented across isolated systems. This fragmentation leads to significant inefficiencies, administrative waste, and suboptimal care. By standardizing data and making it more available and transparent, organizations can enable better communication and collaboration across the entire healthcare ecosystem. This includes combining various data sources, such as price transparency data, claims, clinical information, quality metrics, and CMS benchmarks.
Once all of this data is aggregated and standardized, organizations must invest in a modern technology architecture capable of real-time decision-making, such as application programming interfaces (APIs) and electronic data interchanges (EDIs). By creating this data-driven, interconnected ecosystem, healthcare organizations can establish the necessary foundation for advanced AI-driven solutions that improve affordability, access, and quality of care.
Amit Phadnis, Chief Innovation & Technology Officer at RapidAI
Too often, organizations try to implement AI before they’re ready. The most important step is strengthening the data foundation, ensuring interoperability across systems, quality and diversity of the data sets, normalizing data for consistency, and enforcing strict governance around privacy and security. With those fundamentals in place, health systems are far better equipped to adopt AI solutions that truly improve patient care and operational efficiency.
Michael Mainiero, CIO at Catholic Health
Strong foundations matter more than the flash of new AI tools. Organizations need disciplined master data management, clear governance for use cases, and relentless attention to data quality. AI amplifies whatever you feed it. If the data is inconsistent, the results are too (garbage in, garbage out). Getting the basics right is the single best way to future-proof for AI.
Brian Laberge, Solution Engineer at Wolters Kluwer Health
The biggest consideration for AI readiness is not in the technology itself, but rather the quality of the data feeding into the AI solution. Low-quality data with missing, incomplete, or unverified information can reduce the accuracy of AI tools, in turn limiting the positive payoff they can have for healthcare organizations. With various standards across healthcare systems, organizing different insights across different care settings before feeding the data into an AI solution ensures that it doesn’t lose its meaning and that their model is accurate and impactful enough to better serve both healthcare workers and patients alike.
Kamya Elawadhi, Chief Client Officer at Doceree
From my perspective, AI is only as strong as the data foundation beneath it. For healthcare organizations, preparing for AI means treating data management as a strategic priority rather than a back-office function. That starts with ensuring data quality, eliminating duplication, standardizing formats, and maintaining integrity across various sources like EHRs, claims, and patient-reported inputs.
Equally important is breaking down silos so datasets can actually connect. Fragmented systems limit what AI can achieve, whereas a unified framework allows insights to emerge from the intersection of clinical, commercial, and behavioral data. Compliance and governance are also critical, since even the most advanced models lose value if they cannot operate within regulatory guardrails. To me, readiness for AI is less about chasing the latest tool and more about building an intelligent, system-wide foundation, almost like an operating system, that can scale, adapt, and support innovation without compromising trust.
Matt Donahue, Chief Technology Officer at CloudWave
Healthcare systems operate within some of the most complex, interconnected, and fragmented data environments of any industry. EHRs, imaging systems, medical devices, clinical workflows, billing systems, and compliance platforms all generate and process sensitive protected health information (PHI), much of which is done in real-time. These systems were designed for care delivery, not defense, and often cybersecurity has been layered on after the fact.
Integrating AI into this already complex ecosystem introduces new considerations. Some of the key data management efforts healthcare organizations should undertake to ensure they’re ready for AI solutions involve asking critical questions, such as: how will it interact with different systems? Will it have access to PHI? What models are we using? Where does the data go? Who’s training the model, and on what? Who has access to it, and how is it protected? Is PHI safe? These questions must be answered within the context of strict compliance frameworks, such as HIPAA and HITECH.
To ensure they are ready to adopt AI, hospitals and health systems ultimately need a foundational understanding of what AI can (and can’t) do, and how to apply it responsibly in a healthcare setting. When applied thoughtfully, with the proper safeguards in place, AI can provide hospitals with powerful new capabilities. The key is using AI in the right way, for the right purpose, and in a secure, compliant environment. To help ensure this is done correctly, look for a partner with proven experience in clinical and operational settings, with a strong understanding of workflows, medical devices, patient data, and regulatory requirements. This ensures AI solutions are not only technically sound but also seamlessly integrated into clinical and healthcare IT operations with minimal disruption.
Lani Dornfeld, Healthcare and Business Attorney at Brach Eichler
There are a number of steps healthcare organizations should undertake to ensure they are ready to implement AI solutions. Most importantly is governance and accountability. As with the implementation of most large-scale initiatives, buy-in must start at the top. Depending on the size and resources of a healthcare organization, an AI committee should be formed or responsible individuals should be designated to formulate and assist the organization in developing and implementing organization-wide policies, procedures, and training.
The committee or others responsible for AI integration should establish and implement a process to vet AI vendors and AI tools. Legal or other qualified counsel should be engaged to review AI vendor contracts and terms and conditions of use. Privacy, security, and other compliance officials within the organization should be included in the conversation to ensure regulatory compliance. When possible, the organization should encourage multidisciplinary collaboration, inviting input from different disciplines and capabilities. And, of course, organizations must be prepared to monitor, get organization-wide feedback, and for the continuing changing AI landscape.
Niraj Katwala, Vice President of Engineering at Edifecs, a Cotiviti Business
Healthcare organizations must focus on their existing solutions with AI rather than buying net new. The focus should be on supercharging the in-production solutions they have spent millions building.
AI-Enabling a solution involves making sure that all data that is being fed into the system (input), and that is being generated by the system (output), is being captured accurately and completely. This data is what drives the AI component of these systems. This data needs to be captured and stored in a data warehouse. Finally, you need an ML training infrastructure which can access this data warehouse and build intelligent models and agents that can help improve these applications.
Gary Singh, Senior Director, Product Management at Edifecs, a Cotiviti Business
Healthcare organizations cannot leapfrog to AI-driven insights without first resolving data fragmentation, governance, and quality issues. Once foundational data management is in place, predictive analytics and AI agents can transform population health management, from anticipating risk trajectories to dynamically adjusting value-based care contracts. The winners will be those who treat data as a strategic asset, build trust in AI-enabled decision support, and tightly couple analytics with financial and clinical outcomes.
Building this foundation requires clear governance frameworks, stewardship roles, and harmonization with standards like FHIR and USCDI. Accurate identity management, data provenance tracking, and security safeguards such as tokenization and role-based access are non-negotiable. Continuous monitoring for data drift, bias, and completeness, especially for inconsistently captured social determinants of health data, is critical. Without this groundwork, AI outputs risk being unreliable or inequitable, undermining the transformation these technologies promise.
Such great answers to think about here! Huge thank you to everyone who took the time out of their day to submit a quote to us! And thank you to all of you for taking the time out of your day to read this article! We could not do this without all of your support.
What do you think are some of the key data management efforts healthcare organizations should be doing to ensure they’re ready for AI solutions? Let us know over on social media, we’d love to hear from all of you!
No comments:
Post a Comment