Executive Summary
BritiAI submission — NHS SBS Healthcare AI Solutions Open Framework (SBS10523)
Lots 5, 6 and 7 — Prime
The NHS does not need another AI vendor. It needs an operating partner that treats clinical safety, UK data sovereignty, and measurable outcomes as the starting line rather than the marketing line. BritiAI is bidding into this framework on that basis, and on that basis alone.
BritiAI is a UK-incorporated artificial intelligence company purpose-built to bring frontier AI capability into regulated environments. For this framework we lead as prime contractor with two named, contracted partners: Votee AI, whose MAGIC large-language-model training platform and V-Note on-premises ambient capture engine are in production across enterprise, government and financial services; and Beever AI, Votee’s peer-reviewed research laboratory, whose ACL 2025 publications and open-source Beever Atlas neural memory layer let us ground every technical claim in this submission in a citation rather than a slide. Our partner relationships are governed by a binding Teaming Agreement that flows down every UK GDPR, DSPT and DCB0129 obligation imposed by an Approved Organisation onto every layer of our supply chain.
Three pillars define the BritiAI offer to NHS Approved Organisations.
First — sovereign by design. Every productised solution in this bid can be deployed inside an NHS Trust’s own network boundary, on-premises if the Trust requires it. There is no transit through United States infrastructure, no foreign sub-processing, no model training on Trust data without the Trust’s explicit written instruction. UK data residency is the default state of our stack, not a configuration option. This matters because the dominant AI vendors entering the NHS market today route audio, prompts and embeddings through hyperscaler cloud regions outside the UK, then ask Trusts to accept that as the cost of doing business. BritiAI does not.
Second — clinically safe without overclaiming. Every solution in this submission ships with a Clinical Safety Case aligned to DCB0129:2018, a Hazard Log, a Clinical Risk Management Plan, and the engagement model a Trust’s Clinical Safety Officer needs to discharge their own DCB0160 responsibilities. We have appointed a named Clinical Safety Officer for this framework. We have deliberately scoped our Lot 5, 6 and 7 offers outside Class IIa medical device territory: our systems support clinicians and operational teams, they do not make diagnostic or treatment decisions. Where a future call-off requires a medical-device claim, BritiAI will follow the MHRA software-as-a-medical-device route in full before commencing the work. The NHS has seen too many vendors learn the difference between “AI-assisted” and “AI-decisive” after deployment. BritiAI starts on the right side of that line.
Third — embedded, not delivered. Our Lot 7 advisory wrap is built around a Fractional Chief AI Officer service — senior leadership embedded 1–3 days a week, accountable to the Trust Board for the Trust’s AI portfolio. This is designed for organisations that need leadership before they need licences. It is also our answer to the most-repeated frustration in NHS digital transformation: the gap between the strategy deck and the in-flight programme. BritiAI closes that gap by living inside it.
Across the three lots we are bidding, BritiAI commits in writing to the following headline outcomes, each measured at a Trust-agreed baseline and reported at 90 days from go-live:
- Scribe-On-Site (Lot 6 — Ambient Clinical Documentation): ≥60 minutes per clinician per day of documentation time reclaimed.
- WardFlow Copilot (Lot 6 — Agentic Discharge Orchestration): ≥25% reduction in median discharge-letter cycle time.
- Atlas for Trusts (Lot 6 — Neural Memory Layer): ≥40% reduction in median time-to-find across policy, SOP, clinical guideline and learning-log corpora.
- MAGIC-NHS (Lot 5 — Sovereign LLM Training Pipeline): a benchmarked, model-carded, governance-documented fine-tune delivered to the Trust within an agreed envelope, on infrastructure the Trust controls.
- Trial Concierge (Lot 5 — Agentic Trial Recruitment & Protocol Support): measurable uplift against Trust-set recruitment and protocol-response baselines, with full clinical sign-off at every automated step.
Each commitment is underpinned by an outcome-linked pricing option, in which a portion of fees is held back and released against documented achievement of the KPI. We are willing to put commercial risk where our technical claims sit.
The remainder of this response sets out, lot by lot, how BritiAI proposes to deliver these commitments, the evidence behind them, the governance that surrounds them, and the social value we will generate alongside them. We have written every line of it for a sceptical NHS Chief Clinical Information Officer reading at the end of a long day — not for a marketing audience, and not for ourselves.
BritiAI is asking for a place on this framework so that NHS Approved Organisations can buy frontier AI capability from a UK supplier, deploy it inside their own boundary, hold the supplier to a published outcome, and know that the clinical safety paperwork was written before the demo, not after the incident.
That is the offer. The detail follows.