Scaling AI: What we heard and what we recommend -post NHS Confed 2026

The NHS has the ambition, energy (and now the licences!) to change the way it works  – but good implementation plans are still scarce 

We were impressed at the quality and enthusiasm of the clinical, operational and digital leadership teams we met and the presentations made that talked about how to harness AI to improve patient outcomes, reduce administrative burden and help staff work more effectively.  Especially the 4 clinicians who came over to discuss how to set up local azure-based LLMs. 

Three years into the most disruptive technological innovation in 40 years, the conversation has shifted – the focus has shifted from innovation to execution. The most frequent question we were asked is how best to deploy co-pilot safely, delivering value, and scaling proven, successful approaches across the NHS?  

Our take: the current, biggest National challenge is how best to scale, at pace, the best work happening now across the NHS, so no NHS organisation need to recreate work successfully completed elsewhere.  

This blog outlines some practical steps to help move you forward. 

1. Focus on the value from AI, not just adoption 

What we heard 

Discussion centred on AI-enabled productivity tools, ambient voice technologies, clinical support solutions and generative AI. However, the focus was increasingly on outcomes rather than capability. Boards and leadership teams want to know whether AI is reducing administrative burden, creating capacity, improving staff experience and ultimately contributing to better patient outcomes. 

There was also recognition that access to technology is becoming less of a barrier in some areas. However, access alone does not create value. Organisations must still demonstrate that investments are delivering measurable benefits. 

Our view 

The next phase of AI adoption will be defined by evidence, not enthusiasm. 

Many organisations have already completed pilots or early deployments. The challenge now is proving sustainable operational, clinical and financial impact. As commercial models increasingly become usage-based, understanding whether adoption is actually generating benefits becomes even more important. 

What organisations should do next 

• Focus on a small number of high-value use cases. 

• Measure outcomes, not implementation activity. 

• Link adoption to organisational priorities and benefits. 

• Build business cases based on evidence rather than expectation. 

2. Establish Trust and Clinical Confidence to enable scale at pace 

What we heard 

One of the strongest themes throughout ConfedExpo was trust. 

Importantly, we did not see widespread resistance to AI. Many clinicians and healthcare professionals could clearly see its potential benefits. What we heard instead were reasonable questions around clinical safety, information governance, data sovereignty, reliability, regulatory compliance and appropriate human oversight. 

Our view 

In healthcare, governance should not be viewed as a barrier to adoption. It is what enables adoption at scale. 

No organisation should deploy AI into operational or clinical workflows without confidence in its safety, effectiveness and accountability. The organisations making the strongest progress are those embedding governance, assurance and clinical engagement from the outset, rather than treating them as separate activities. 

Trust is built through evidence, transparency and clear accountability. 

What organisations should do next 

• Establish governance and assurance frameworks early. 

• Engage clinicians and users early (expect some to ask difficult questions) and throughout deployment. 

• Define appropriate use cases (there are plenty already proven) within clear boundaries. 

• Build confidence through evidence and communication. 

3. Scale What is Already Working 

What we heard 

Perhaps the most consistent message was the challenge of moving from successful pilots to organisation-wide impact. 

Across trusts and wider NHS organisations, common themes emerged: change management, governance, financial pressures, demonstrating value and sustaining adoption at scale. 

Our view 

As evidence of AI deployments grows exponentially across the NHS, organisations should establish a key principle – wherever we begin with co-pilot, make sure there is a clear plan to scale from the start.  

There will always be cases where local validation is needed. However, many NHS organisations are now facing similar challenges and pursuing similar objectives. The opportunity is to learn from proven deployments and accelerate adoption rather than repeatedly proving the same concepts. 

This is particularly important in a financially constrained environment. For many organisations, the greatest opportunity lies not in new technology investments but in maximising value from tools already deployed and embedding them into everyday workflows. 

What organisations should do next 

• Scale proven approaches where evidence already exists. 

• Learn from successful NHS deployments. 

• Invest in adoption and change management. 

• Maximise value from existing technology investments. 

• Establish clear executive ownership for outcomes. 

4. Impact 

Accelerated scale deployment, using the approaches and actions outline above will bring £10,000s of productivity benefits forward and into FY26/27. 

5. Conclusion  

The volume co-pilot purchase could really could be different from previous digital disruptors and hugely helpful. But (and here is the plot twist) harnessing AI technology to accelerate collaborative working and implementation will be as important as delivering it well to the frontline.