New survey reveals where AI is moving the needle in healthcare in 2025

By Nate MacLeitch, CEO of Quickblox

In our latest survey, “AI Tool Adoption Healthcare (2025),” Quickblox discovered healthcare providers are moving from pilots to purpose when it comes to AI.

We surveyed 101 healthcare leaders to determine where AI is taking hold and where it’s hitting resistance.

Responses came from a wide cross-section of the healthcare sector, led by hospitals (32.7 per cent), clinics (21.8 per cent), and healthtech startups (16.8 per cent), with over half representing large organisations.

 

The findings show that AI use is no longer about experimentation alone; it’s about deployment with intent, especially in areas that relieve everyday work and improve access to care.

The majority cited operational efficiency (81 per cent) as their top reason to adopt AI, closely followed by cost reduction (65 per cent) and improved patient care (63 per cent) – indicating that healthcare providers are starting with inward-facing efficiencies and are prepared to extend AI to patient-facing domains.

Quickblox’s latest research comes at a time when more than 70 per cent of healthcare professionals are already pursuing or implementing generative AI initiatives.

Most early efforts focused on automating administrative workflows and augmenting workforce productivity rather than clinical decision-making.

Together, these findings reveal a healthcare sector moving past curiosity into practical adoption that focuses on low-risk use cases that will build the foundations for broader AI integration.

Momentum starts with healthcare workflows

In the latest Quickblox data, administrative and operational functions dominate the early wave of adoption: 68 per cent said operational efficiency, and 62 per cent said administrative tasks are where AI can have the most impact.

By contrast, only about 37 per cent see a significant impact on diagnosis.

That 30-point differential signals that organizations are prioritizing the low-risk, high-return tasks first, while they build the trust, governance and infrastructure required for more sensitive clinical use cases.

When respondents answered which AI solutions they were most interested in, tasks such as documentation assistance, scheduling, virtual assistants, and administrative workflow oversight were where organisations expected the most impact and saw the most activity.

In contrast, clinical applications such as diagnosis and triage are emerging but still trailing:

• Workflow automation tools – 74.3 per cent
• Patient engagement platforms – 47.5 per cent
• Virtual assistants – 47.5 per cent
• Diagnostic support tools – 46.5 per cent
• Predictive analytics – 45.5 per cent

McKinsey’s survey, reported by HIT Consultant, echoed the focus on lower-risk use cases.

It revealed that while 85 per cent of healthcare respondents are exploring or using generative AI, many remain in proof-of-concept mode for higher-risk clinical functions.

Outside of the survey, Quickblox sees a similar pattern with healthcare providers adopting a “crawl-walk-run” pattern.

Often, our clients start with simple chat features on their practice websites before expanding to features such as AI intake.

Security, integration, and trust make the triad of AI adoption

Data privacy and security top the barrier list in Quickblox’s study, with 45 per cent of respondents concerned about governing personal health information. Integration complexity and cost follow next.

This reflects a sector that remains cautious about protecting sensitive data, even as confidence in AI’s potential grows.

Yet interestingly, despite the concern, over 59 per cent say they are confident in their data governance abilities.

That suggests healthcare professionals view privacy as a manageable challenge, not a blocker.

With doubts around integration complexity and data privacy top of mind, vendor alignment with long-established compliance frameworks like HIPAA remains central to earning credibility with healthcare organizations.

The World Economic Forum’s “Future of AI-Enabled Health” 2025 report supports this statement, calling for stronger governance, data alignment, and ethical frameworks as AI scales in healthcare.

Meanwhile, McKinsey’s general AI report notes only 1 per cent of organisations consider themselves fully mature in AI deployment, underscoring that governance and workflow redesign remain critical to achieving consistent, scalable AI adoption.

Together, these findings indicate that the technology is ready, but ease of integration and trust are still the differentiators.

As those foundations strengthen, investment is accelerating; Quickblox found that 71 per cent of healthcare respondents plan to invest in AI in the next 12 months.

Yet widespread deployment does not automatically equal scaled impact.

Organisations that leap too quickly into high-risk clinical use cases without foundational readiness may stumble.

The promise of AI is unraveling, but success relies on people, process, and readiness, not just technology.

Done right, healthcare providers can move from “thinking about AI” to “deriving tangible value from AI” and, ultimately, improve patient outcomes.