AI is arriving in care providers faster than most organisations can redesign around it. The opportunity is compelling: better care, more capacity and a health system that can adapt as demand changes. But there is a catch. If staff still have to hunt for information, check every output and move it between disconnected systems, AI has not transformed the pathway — it has simply created another task. The winners will not be the providers with the most AI. They will be the ones that make care work materially better because of it. That starts with three deceptively simple moves:
This blog was inspired by Srikanth Sankaran Iyer’s article, From tool sprawl to agentic commercial operations1. His central challenge travels well into care delivery: organisations do not have an AI use-case problem; they have an orchestration problem. Pilots, copilots, predictive models and specialist tools are multiplying. For provider leaders, the harder and far more valuable job is to connect them with trusted data, real clinical workflows and human expertise in one governed operating model. Otherwise, experimentation grows faster than value.
69% of healthcare and life-sciences organisations surveyed were using generative AI or large language models in 2026, up from 54% a year earlier1
Picture a familiar scenario. A new AI tool produces a report in seconds, but a clinician still has to copy it into another system, reconcile it with the patient record and decide whether it can be trusted. The demo looks impressive; the working day does not improve. Replicate this across radiology, outpatients, theatres, discharge and administration, and providers can end up with more pilots, more checking and more cost, but little movement in patient flow, staff experience or outcomes.
68% of 2,041 healthcare leaders surveyed across 90 countries did not feel very confident using or evaluating AI tools1
For care providers, this is not a theoretical risk. Services already run across fragmented records, ageing infrastructure, organisational boundaries and pathways built on dozens of hand-offs. AI can help, but only when leaders start with how care is delivered, not with what technology can do. The test is not whether an individual tool works. It is whether patients move through the pathway more quickly, staff spend less time compensating for broken processes and decisions become safer and more consistent.
|
AI without orchestration
More tools can mean more complexity—not better care.
|
|
|---|---|
|
1. Fragmented technology Separate tools can reproduce the silos already beneath them. |
2. Workflow complexity Improving one task does not automatically improve the whole pathway. |
|
3. Poor information Incomplete and disconnected data limits safe, effective adoption. |
4. Human oversight Clinical judgement, validation and accountability must remain visible. |
|
5. Added workload Checking and moving outputs can swallow the productivity gain. |
6. Operating-model gaps Ownership, monitoring, permissions and escalation need clear answers. |
Resist the temptation to start with another technology project. Start with a pathway where outcomes, access, productivity or experience are visibly under pressure. Walk the work from end to end: where information is created, where decisions stall, where delays build and where staff bridge gaps between systems by hand. Then ask where AI can genuinely remove friction, support judgement and improve care.
Build governance into the pathway from the start. Be clear about clinical ownership, the limits of autonomy, the data the AI can use, when people must intervene and how concerns will be escalated. Above all, measure what leaders and teams actually care about: time to diagnosis, avoidable hand-offs, clinician time released, patient experience and outcomes. Count improvements in care, not the number of pilots launched.
Get this right and the prize is not incremental. It is better, more consistent care; meaningful capacity from scarce resources; and an organisation that can respond faster as demand, workforce pressures and treatments change. AI stops being another deployment and becomes part of how the provider learns and improves every day.
When AI is embedded across the pathway, it can turn fragmented information into timely, useful insight, right where clinical and operational teams need it.
The benefit: better decisions, more consistent care and the potential for improved outcomes and experience.
This may be the biggest economic prize. Redesigning workflows so AI, analytics and existing systems work together can reduce duplication, manual hand-offs and the burden of people acting as the integration layer.
The benefit: more clinical and operational capacity without matching growth in workforce or infrastructure.
Connect data, AI, workflows and governance, and providers are better placed to respond to changing demand, workforce pressure, new treatments and new models of care. Intelligence becomes part of everyday operations, not a fresh technology project every time something changes.
The benefit: a health system that can learn, adapt and redesign care more quickly.
Examples:
More than one hour earlier
A national study covering 452,952 stroke admissions across all 107 NHS hospitals admitting acute stroke patients in England found that AI-supported sites doubled thrombectomy rates from 2.3% to 4.6%3
43 minutes per person per day or more
Following a trial involving more than 30,000 staff across 90 organisations, the NHS announced rollout of Microsoft 365 Copilot to 505,000 clinicians and support staff in 20264
Taken together, these benefits reinforce one another: better information and decisions improve care, redesigned workflows release capacity, and both make the organisation more able to adapt. The graphic below summarises that prize and why provider leaders should judge AI by the improvement it creates across the system, not by the technology alone.
| BETTER CARE | MORE CAPACITY | GREATER AGILITY |
|---|---|---|
| Better decisions, more consistent care and improved experience. | Less duplication and more value from scarce people and assets. | A system that learns, adapts and redesigns care more quickly. |
This is where IQVIA can help: not by starting with a technology catalogue, but by helping provider leaders turn a pressing pathway problem into measurable improvement. We connect clinical and operational strategy with healthcare data, analytics, digital infrastructure, governance and delivery, so AI is shaped around the service, the workforce and the patient.
That means choosing the right pathway, building the evidence and data foundations, redesigning the work with clinical teams, defining safe oversight and following benefits through to delivery. The objective is simple to state but hard to achieve: better decisions, released capacity and safer, more consistent care at scale.
Three steps to take now:
So the question for provider leaders is not, “Where can we deploy AI?” It is, “Which pathway must work materially better and what would we have to change to make that happen?” Start there, and AI has a chance to become something far more useful than another pilot.
1. https://www.iqvia.com/blogs/2026/09/from-tool-sprawl-to-agentic-commercial-operations
2. NVIDIA (2026) State of AI in Healthcare and Life Sciences: 2026 Trends. Accessed 28 September 2026; Miller, I.
3. (2025) ‘Life-changing’ AI support helping stroke patients get a second chance, 2 December. Accessed 28 September 2026; NHS England
4. (2026) 500,000 NHS staff to get new artificial intelligence tools to help free up more time for patients, 8 June. Accessed 28 September 2026.
5. (2026) 3 Key Insights for the 2026 Health AI Horizon. Digital Medicine Society, 29 January. Accessed 28 September 2026; NHS England
