Ayoob AI

AI Automation for Leeds NHS Trusts and Clinical-Letter Triage

·5 min read·Husain Ayoob
AI automationLeedsNHShealthcareclinical lettersUK business

Leeds is the NHS digital capital of England. NHS Digital is headquartered in Leeds. Leeds Teaching Hospitals NHS Trust is the largest in the country by patient throughput. The West Yorkshire health-tech cluster around the city is the densest concentration of NHS-adjacent technology suppliers outside London. For an AI agency, that is a high-stakes, slow-moving, deeply regulated procurement environment. For the Trusts and the suppliers operating in it, the question is what to put AI on without ending up in front of an information governance committee defending a data export you should not have made.

This is what we build into Leeds NHS environments.

The Leeds NHS automation workload

Three high-value workloads recur across Leeds NHS engagements.

Clinical-letter triage and referral routing. Consultant clinics across Leeds Teaching Hospitals, Mid Yorkshire, and the wider West Yorkshire ICS handle volumes of incoming letters and referrals that take consultant or specialist nursing time to read, classify, and route. The cognitive load is not the clinical decision; it is the working-through. AI that reads, classifies by specialty and urgency, and routes with full audit context removes that load while keeping the clinician in the decision authority.

Subject Access Request processing. GDPR SAR turnaround inside the statutory window is a recurring information governance challenge for any NHS Trust at scale. The work is search across multiple clinical and operational systems, redaction of third-party identifiers and exempt material, response packaging, and audit. We have shipped SAR pipelines that compress 22-day turnarounds to under 4 hours of working time, with the IG officer signing off.

Research-data governance. Leeds NHS Trusts host substantial research activity in collaboration with the University of Leeds, the AI Hub, and external partners. Managing research-data flows under HRA, REC, and patient-consent constraints is a documentation-intensive process. Full code AI that maintains the data flow records, consent register, and access audit trail removes a recurring administrative cost from research delivery.

Why no third-party LLM

The hard architectural constraint for NHS workloads is that patient identifiable data does not leave the controller without explicit lawful basis and patient awareness. The standard commercial LLM contract sends content to the provider for inference. That is a data export. For NHS use, it is not an acceptable architecture.

Full code AI deployed inside the Trust's own infrastructure removes the constraint entirely. The model runs in the Trust's tenancy. Inference happens on infrastructure the Trust controls. The audit trail is in the Trust's own logging stack. No data leaves.

This is also the architecture that survives an NHS DSPT review without compromise. ISO 27001:2022, Cyber Essentials, ICO-registered, GDPR-compliant by design, with the technical evidence the IG officer needs for the toolkit submission.

How we deliver into Leeds

Ayoob AI is registered at Newbridge Street, Newcastle upon Tyne. Newcastle to Leeds is a 90 minute direct train, and Leeds is a regular part of our delivery footprint.

Discovery is in person, in Leeds, in week one. Technical design review at month one is in person. For Trust engagements that need weekly on-site presence, particularly during information governance reviews, clinical safety case compilation, and NHS DSPT evidence gathering, that is part of the retainer scope at no extra cost.

The procurement profile fits NHS direct award. Place on CCS RM6200 (AI Dynamic Purchasing System) and RM6173 (Automation Marketplace), so Trust procurement teams can direct-award without re-tender where the framework call-off conditions permit. Five pending UK patents on the GPU and AI compute infrastructure underneath the work.

Clinical safety and the Trust's CSO

We do not own the clinical safety case. The Trust's clinical safety officer does. Our role is to make compiling DCB0129 and DCB0160 evidence straightforward rather than a multi-month archaeology project. We provide the technical artefacts, audit trails, and documentation the CSO needs. The CSO signs.

The boundary matters. Suppliers who claim to deliver a CSO-ready system without the Trust's CSO involved are usually misunderstanding the regulation. We take the boundary seriously because it is the only structure that survives a serious review.

Pricing and commercial shape

Existing systems retainer from £4,000 per month. New systems retainer from £6,000 per month. Both on a 12-month minimum term. Exact pricing is set on consultation against a written scope, fixed inside one week of the discovery call.

Hosting and model API costs sit outside the retainer and are paid by the Trust to their own cloud and model providers. No per-seat software licences. No marked-up cloud costs. No third-party LLM subscription. The retainer commercial model is covered in detail at what AI automation actually costs a Newcastle SMB, and the broader Leeds delivery picture is at AI automation Leeds.

Getting started

The first step is a 30 minute discovery call. We tell you straight whether the work is a fit, and if it is we send a written scope and a fixed monthly number within a week.

Book a discovery call.

About the author
Husain Ayoob, Founder & CEO, Ayoob AI Ltd
Husain Ayoob

Founder & CEO, Ayoob AI Ltd

BSc Computer Science with AI, Northumbria University 2024. 5 UK patents pending covering the Ayoob AI stack. ISO 27001:2022 certified (organisation).

Full bio, patents, and press →

Frequently asked questions

Can AI handle NHS clinical-letter triage in a way that is actually safe?

Yes, with the right architecture. The system does not make clinical decisions. It triages, classifies, and routes letters and referrals to the right specialist with the right urgency band, with full audit traceability. The clinician retains the decision authority. The AI removes the cognitive load of working through the letter pile, which today often runs to days of consultant time per clinic per week. Architecturally, the model runs on the Trust's own infrastructure, no patient data leaves the Trust, and every decision the AI takes is logged with the source content for audit. That is the model that survives an NHS DSPT review.

Is third-party LLM acceptable for NHS clinical workloads?

Not at the point of patient identifiable data. NHS Digital, ICO, and the Caldicott Principles are clear that patient data leaving the controller without explicit lawful basis and patient awareness is not acceptable. The standard third-party LLM contract sends content to the provider for inference, which is a data export. For NHS use, the AI must run inside the Trust's tenancy, on infrastructure the Trust controls, with no third-party LLM provider in the data path. That is what we build.

What does SAR automation cover?

End-to-end automation of Subject Access Request handling against the GDPR statutory deadline. Receipt and validation of the request, identification of relevant records across the Trust's clinical and operational systems, redaction of third-party identifiers and exempt material, packaging of the response, and audit trail. We have shipped SAR pipelines that compress 22-day turnarounds to under 4 hours of clinical and information-governance time. The reduction is not in the work; it is in the manual time absorbed by the work. The information governance officer still signs off.

How does Ayoob AI deliver into Leeds NHS Trusts from Newcastle?

Newcastle to Leeds is a 90 minute direct train. We run discovery in person, technical design review in person, and biweekly sprint demos as the project needs. For Trust engagements that require weekly on-site presence (information governance, clinical safety reviews, NHS DSPT evidence gathering), that is part of the retainer scope. Day-to-day delivery is remote-first within the Trust's secure network.

What does it cost?

Existing systems retainer from £4,000 per month. New systems retainer from £6,000 per month. Both on a 12-month minimum term. Exact pricing is set on consultation against a written scope, fixed inside one week of the discovery call. Hosting and model API costs sit outside the retainer and are paid by the Trust to their own cloud and model providers. For NHS direct award, we work under CCS RM6200 (AI Dynamic Purchasing System) and RM6173 (Automation Marketplace).

Is the system DCB0129 / DCB0160 / clinical safety compliant?

Architectures are designed to support the clinical safety case rather than create one to defend. We provide the technical artefacts, audit trails, and documentation necessary for the Trust's clinical safety officer to compile DCB0129 and DCB0160 evidence. The clinical safety case remains the responsibility of the Trust's CSO. Our role is to make compiling that case straightforward rather than impossible.

Want to discuss how this applies to your business?

Book a Discovery Call