The pros and cons of AI in healthcare
What payers need to know before deploying AI in provider navigation
Payers are facing a critical question: what are the pros and cons of AI in healthcare and where does it actually create value vs. introduce unnecessary business risk?
That question is especially urgent in provider navigation, where search and scheduling sit at the intersection of cost, network performance, member satisfaction, and competitive differentiation.
When navigation fails, the consequences aren’t limited to a frustrating digital experience. Payers see more out-of-network utilization, more leakage, lower member trust in the provider directory, and weaker performance on access-related experience measures.
Payers deploying AI in navigation are already seeing the difference in call deflection, in-network rates, and in quality scores. The healthcare AI market was estimated to be over $36B in 2025 ¹ , and 92% of healthcare executives believe AI will provide a competitive edge. ² These are navigation outcomes, and in most traditional nurse line models, they go uncaptured.
Further, it’s estimated that AI could generate $200 billion to $360 billion in net healthcare savings annually, including non-clinical workflows such as scheduling and coordination.³
The upside of AI in healthcare navigation is clear, but there are legitimate tradeoffs that business leaders must consider. Payers that don’t have well-thought AI governance in place and clear methods for de-risking AI systems jeopardize their investment in AI solutions because they will fail even initial user testing, let alone getting to scale.
Here’s a quick summary of the pros and cons of AI in healthcare navigation:
Pros
Scales navigation without increasing costs
Improves in-network utilization
Expands access beyond business hours
Directly impacts quality and experience metrics
Cons
Governance failures create legal and reputational risk
Poor data quality undermines outcomes
Creates fragmentation if not integrated end-to-end
Costs of building and maintaining AI are widely underestimated
The pros of AI in healthcare
1. Streamlines and changes the cost structure of navigation
Provider navigation is administratively intensive. Members need to manually find in-network providers, confirm availability, schedule appointments, and navigate follow-up. Traditionally, this complexity drives cost, with more demand meaning more staff. AI changes this equation.
AI can automate this orchestration at scale, reduce the cost per interaction, and handle high-volume interactions without proportionally increasing operational costs.
2. Improved in-network utilization and cost control
Out-of-network care can cost payers significantly more than equivalent in-network services. When navigation breaks down, members default to options that are most accessible, often outside the contracted network. AI-powered navigation guides members to the right in-network provider at the right time, directly improving network utilization, reducing downstream costs, and protecting the plan’s financial performance.
3. Always-on availability
Members need care on evenings, weekends, and outside business hours, and not just when a call center is staffed. AI-enabled navigation provides 24/7 access to provider search, scheduling support, and care guidance, reducing both member frustration and unnecessary utilization of high-cost settings like ED or urgent care.
4. Impact on quality metrics
Faster, more accurate navigation can lift CAHPS scores, HEDIS measures, and Star ratings, which are directly tied to access and timeliness of care. When AI follows up to close care gaps, the link between navigation performance and revenue becomes measurable.
The cons of AI in healthcare
1. Ungoverned AI underperforms and creates liability
High-profile AI failures at major health insurers have made the regulatory and reputational risks of poorly governed AI concrete.
This isn’t a niche legal concern. Gartner estimates that by 2027, 60% of organizations will fail to realize anticipated value from their AI use cases due to inadequate ethical governance frameworks.⁴ For payers, the downside of poor governance also includes regulatory action, litigation, and member trust erosion in addition to the financial impact.
2. Bad data in, worse outcomes out
AI navigation is only as good as the data it runs on. Provider directories remain one of the most persistent data quality problems in health insurance, and some reports illustrate as much as 40% inaccuracies in data as simple as phone numbers.
Federal law mandates verification every 90 days, yet errors remain widespread.⁵ Last year, the California Attorney General secured a $40 million settlement with Health Net over allegations of inaccurate mental health provider directories. With No Surprises Act disputes up 119%, the regulatory cost of directory failures is rising fast.⁶
Routing members to providers who have moved, retired, or stopped accepting patients denies care, damages trust, and carries real regulatory risk under the No Surprises Act.
3. AI pilot purgatory—hidden risks, complexities, and costs
One of the most common patterns in healthcare AI failures is the disconnect between what works in a controlled demonstration and what works at scale in a production environment.⁷
In production, data is incomplete and inconsistent, edge cases are the norm, and errors carry clinical, financial, and operational risk that a demo environment simply cannot surface. This is why AI pilots stall and costly rebuilds happen.
4. Hidden cost of building
The appeal of building is clear: full control, proprietary data, and differentiation. But Gartner research has found that agentic AI transformation costs can run several times higher than traditional Enterprise Resource Planning software implementations, with organizations regularly underestimating what’s required across development, integration, governance, data management, and organizational change.⁸
Failed AI projects in healthcare average $6.8M in investment against just $1.9M in delivered value, with most failures driven not by model performance but by system design, fragmented data, and governance gaps that only surface at scale.⁹
What separates effective AI-powered navigation from costly experiments?
AI in healthcare navigation that works in the real world comes down to a few non-negotiable requirements. These are especially critical in provider navigation, where accuracy, timing, and trust directly impact access, cost, and member experience.
Clinical safety and guardrails: AI must detect high-risk scenarios, escalate to human support when needed, and handle uncertainty without fabricating answers.
Real-time data integration: Navigation is only as effective as the data behind it. In provider navigation, that means moving beyond outdated directories to real-time availability. AI agents that actively call providers to verify openings and confirm scheduling overcome the ghost network problem without requiring EHR integration.
Workflow orchestration: Effective navigation manages the entire care access journey (search, scheduling, follow-up) within a single workflow integrated into the payer experience.
Monitoring and governance: At scale, you cannot review every interaction manually. Production-grade healthcare AI needs automated monitoring, alerting, and escalation protocols, built in from inception and not retrofitted after a failure.
AI that doesn’t earn the trust of members, pass regulatory scrutiny, and improve real outcomes leads to risks, and not innovation.
The strategic question for payers
The pros and cons of AI in healthcare for provider navigation ultimately comes down to a few strategic considerations: can your organization deploy AI safely, integrate it into real workflows, and maintain clinical and operational accountability at scale?
Payers that have tried to build this capability from scratch often find themselves several years and millions of dollars into a project that’s still not ready for production. Generic AI tools aren’t trained on health plan data, don’t understand network logic, and aren’t designed for clinical accountability. The organizations that are moving fastest aren’t just adopting AI; they’re partnering with platforms purpose-built for the stakes of healthcare. That distinction matters more than which model is under the hood.
Explore how Pager Health Provider Navigator delivers AI-first, human-always navigation, from provider search and real-time scheduling to follow-up care, fully integrated into the existing payer experience.
Connect with Pager Health to explore how AI-powered orchestration can help your plan simplify member navigation, improve engagement, and deliver measurable ROI.