Barriers to Scaling AI Initiatives and Effective Strategies to Overcome Them

We talk a lot about AI in a theoretical sense – how could AI improve this, what cybersecurity should we plan for AI, etc. But what is the reality of the situation? Today, we are seeking out the first-hand, personal experiences people have had with implementing AI in healthcare! We reached out to our incredible Healthcare IT Today Community to ask: what barriers – technical, cultural, or financial – have you encountered when scaling AI initiatives, and what strategies have proven most effective in overcoming them? Below are their responses.

Harshit Jain, Founder and Global CEO at Doceree

Scaling AI in healthcare isn’t just a technology challenge—it’s an ecosystem one. Technical barriers include fragmented data and limited interoperability. We address this through standardized data frameworks and deeper integration with existing clinical systems like EHRs. Cultural resistance stems from clinician skepticism. Building trust requires transparency, clear use cases, and keeping humans in control—not replacing clinical judgment but augmenting it.

Financial concerns around ROI are real. Organizations need proof. Starting with targeted, high-impact use cases that demonstrate measurable improvements in efficiency or patient outcomes builds momentum and justifies broader investment. The most effective strategy? Align AI initiatives with actual clinical and operational needs—not innovation for its own sake. When AI solves real problems providers face daily, adoption follows naturally.

Dr. Scott Schell, Chief Medical Officer at Cognizant

Algorithm performance is rarely the limiter. Barriers fall into three buckets:

Data Interoperability Gaps: Poor data quality, fragmented records, and siloed systems undermine reliable performance and integration

Workflow Misalignment: If tools are not embedded in core clinical and administrative systems, they will tend to be ignored or abandoned

Organizational Readiness: Lack of governance structures, unclear ownership of outcomes, and limited operational capacity prevent scaling beyond pilots

Denis Whelan, CEO at Documo

The biggest barriers we’ve seen when scaling AI initiatives are cultural resistance to change, disruptions to existing workflows, and uncertainty around ROI. Teams overcome these challenges by embedding AI directly into the systems staff already use, starting with low-risk workflows where impact can be measured quickly, and focusing on tangible improvements – like reducing repetitive manual tasks, errors, and delays in treatment – rather than positioning AI as a replacement for staff. Demonstrating early wins builds trust, encourages adoption, and creates momentum for scaling more complex initiatives.

Jitin Asnaani, Chief Product Officer at Rhapsody

AI pilots are easy. Enterprise AI is hard, and healthcare is about to feel that difference. Scaling AI is less about ‘choosing the right model’ and more about removing the bottlenecks that stop it from operating in the real world. Disconnected, siloed systems, duplicated identity data, and slow security approvals keep AI stuck in pilot mode. The organizations that scale will standardize data access through proven systems of connectivity, strengthened governance, and embedded AI into existing workflows, so teams spend less time stitching data together and more time driving measurable outcomes.

Ryan Hungate, DDS, MS, Chief Clinical and Strategy Officer at Henry Schein One

Scaling AI is far less about the technology itself and much more about trust, workflow integration, and measurable value. Technically, the biggest barrier is rarely model capability—it’s data quality, interoperability, and embedding intelligence directly into the clinical and operational moments where decisions are made.

Culturally, adoption only happens when teams believe AI is augmenting their professional expertise rather than replacing it, which means transparency, clinical validation, and clear accountability are essential. Financially, organizations struggle when AI is positioned as experimentation instead of transformation; the breakthrough comes when outcomes are tied directly to productivity, revenue cycle performance, or improved patient experience.

Ben Hilmes, CEO at Healthcare IT Leaders

The biggest barrier to AI adoption isn’t technical; it’s cultural. Clinicians have been burned by technology that promised to simplify their lives but instead increased documentation burdens. To overcome this, we must involve frontline clinicians from day one, deliver quick wins that genuinely reduce workloads, and demonstrate a willingness to pivot if a tool isn’t performing. Adoption is earned through transparency, not mandated by decree.

Generic AI solutions rarely succeed without adaptation. What works in an academic medical center may fail in a rural hospital because patient populations and workflows are fundamentally different. These nuances are why embedded leadership is essential; you need leaders who understand not just the technology’s potential, but how it must be configured to fit the specific realities of local care delivery.

Gokul Mohan, CEO at CareHarmony

The biggest barriers are cultural, not technical. Many clinicians have been burned by tools that promised efficiency and delivered more work.

What works is showing, not telling. When AI clearly reduces documentation, prioritizes outreach, or prevents issues from escalating, adoption follows naturally. Financial alignment also matters. AI initiatives are much easier to scale when they support value-based care models and outcomes that organizations are already accountable for.

Amy Rettler, SVP of Client Partnerships at Evergreen Healthcare Partners

After assessing our client base, we noted that when it comes to clinical AI, roughly 30-35% of initiatives focused on direct clinical care, such as ambient listening, prioritization, risk identification, and decision support, rather than autonomous decisions. The constraint isn’t the technology; it’s trust. When clinicians lack clarity into how recommendations are generated or where accountability sits, adoption stalls.

Beth Godsey, General Manager of Tendo Insights at Tendo

A common technical barrier is fragmented data and variability in workflows, which can limit how actionable AI insights are. Culturally, frontline teams are often skeptical of tools that add noise or feel disconnected from daily decision-making. Financially, leaders may struggle to justify investment without a clear operational impact.

The most effective strategies focus on embedding AI into existing workflows, co-designing solutions with clinicians and operators, and targeting use cases tied directly to capacity constraints, staff workload, or patient safety risks. When AI clearly helps teams prioritize work and reduce friction, adoption and scale tend to follow.

Lesley Berkeyheiser, Senior Director of Accreditation Strategy and Development, CCSFP at DirectTrust

One of the most difficult challenges with AI implementation is that its impact is enterprise-wide and evolving at an extremely rapid pace. This makes a flexible governance model essential, one that requires organizations to continuously evaluate how artificial intelligence is being used and the risks it poses to systems and data.

Following established standards such as the National Institute of Standards and Technology (NIST) AI Risk Management Framework provides a methodical approach to assessing artificial intelligence tools from initial evaluation through ongoing maintenance, with ongoing risk assessment at the core. At the same time, AI adoption is accelerating as workforce members already use these tools in their personal lives, making “shadow artificial intelligence” a growing concern.

Organizations must proactively educate staff on what is and is not permitted so artificial intelligence can be used safely and securely while still delivering the efficiency these tools provide.

Jackie Mattingly, Senior Director of Consulting Services at Clearwater Security

The biggest barriers to scaling AI in healthcare aren’t just technical; they’re about capacity, resources, and trust. Large systems struggle with data governance and cross-department coordination, while small and rural hospitals often lack the time, budget, and internal expertise to even safely evaluate AI, which can create real fear about losing the human element of care or replacing clinicians.

What’s helped is reframing AI as a way to reduce burden and burnout, focusing on clear use cases that solve real daily problems, involving frontline staff early, and putting basic guardrails in place now rather than waiting for formal regulations to catch up.

Firoze Lafeer, SVP of Data Engineering at Revecore

Scaling AI in healthcare presents a distinct set of challenges across technical, cultural, and financial dimensions.

From a technical standpoint, data quality and infrastructure are often the most significant barriers. Many organizations rely on legacy systems and fragmented data sources, limiting the effectiveness of advanced analytics. Investing in modern, scalable data architectures is a foundational step—not only for AI adoption, but for strengthening the organization’s overall information ecosystem.

Cultural barriers can be just as significant. Skepticism toward automated systems and concerns about reliability or loss of control can slow adoption. Clear communication about the role of AI, combined with well-scoped proof-of-concept initiatives, can help demonstrate value and build confidence. Involving clinical and administrative stakeholders early in the process fosters ownership and helps translate early successes into broader adoption.