BritiAI — NHS SBS Healthcare AI Solutions Framework
Strategic Plan & Recommendation
Prepared for: BritiAI leadership + Votee + Beever Plan baseline: 15 June 2026 — current days-remaining on the bid dashboard Submission deadline: 12:00 noon, Tuesday 21 July 2026 (SAP Ariba) — official ITT deadline Framework Start Date: 26 May 2027 → ~11 May 2035 (8 years) Framework ceiling: £900m incl. VAT Awards expected: March 2027
⏱ COUNTDOWN — REBASELINED TO ITT AMENDMENT (3-WEEK PLAN)
| Plan baseline | 15 Jun 2026 |
| Submission deadline (official) | Tue 21 Jul 2026, 12:00 noon BST |
| Days remaining | See the live bid dashboard |
| Internal submit-by (early target) | Fri 3 Jul 2026, 14:00 BST — submit well ahead of the official deadline |
| Hard backstop | Mon 6 Jul 2026, 14:00 BST |
| Enquiry deadline | 12 Jun 2026 — passed (informal route only via NHS SBS mailbox) |
| Award decision | 19 Mar 2027 |
| Next re-opening | Estimated 12 May 2030 |
Reality check. The official ITT deadline is 12:00 noon, Tuesday 21 July 2026 (Bidders’ Instructions V3). We retain an early internal submit-by of Fri 3 Jul, giving a generous pre-deadline buffer. We have well over three weeks. CSO engagement runs as a standard 5–10-day procurement, not a crisis. CE+ assessment, ISO 27001 Stage 1 documentation, NHS reference letter campaign, and a proper red-team day all comfortably in scope. What this also changes: Lot 8 (Combined Solutions) is automatic — awarded to any supplier winning ≥2 lots. No separate Lot 8 bid. We bid Lots 5, 6, 7 as primary and 3, 4 as stretch (per Wed 24 Jun decision).
1. Executive Summary
NHS Shared Business Services (NHS SBS) has published the Healthcare AI Solutions Open Framework (SBS10523) — the largest dedicated AI procurement vehicle in NHS history. It runs from 26 May 2027 to 11 May 2035, ceilings at £900m incl. VAT (£750m ex VAT), and crucially permits direct award by Approved Organisations to framework suppliers without further competition. Lot 6 evaluation is Quality 90% / Social Value 10% — no cost weighting, which is the rarest possible procurement signal: this is a quality-led race, not a price war.
We mobilised on 15 Jun 2026 against an official submission deadline of Tue 21 Jul 2026, 12:00 noon — with an early internal submit-by of Fri 3 Jul to leave a generous buffer. The plan is a structured programme with proper compliance work, red-team and CEO sign-off cycles. This is achievable but only with: (a) a ruthless lot-selection decision in week 1; (b) a parallel-track readiness sprint on DTAC v2, DCB0129, ISO 27001 and Cyber Essentials Plus evidence; and (c) a defensible consortium structure with Votee (IP/research) and Beever (research lab).
Recommendation in one line: BritiAI bids as prime on Lot 5 (Research, Innovation & Development), Lot 6 (Operational Efficiency) and Lot 7 (Advisory & Specialised Support) — with Votee as named technology partner (LLM, agentic, ambient capture IP) and Beever as the named research/evaluation partner. Lot 8 (Combined Solutions) is automatic if we win ≥2 lots — no separate bid required. Lots 3 (Virtual & Robotic Health) and 4 (Predictive Analytics) are stretch lots, decided Wed 24 Jun if evidence allows.
This positions us where we can actually win (operational/admin/agentic AI is the fastest deployable, lowest clinical-risk class), gives us a route to the higher-margin Lot 8 which is reserved for suppliers awarded ≥2 other lots, and uses Votee/Beever’s strongest assets (MAGIC LLM training, V-Note ambient transcription, Beever Atlas neural memory) without forcing them into a regulated medical device claim they cannot yet evidence.
2. The Framework — What We Actually Have to Win
Source: NHS SBS notice (11 May 2026), Find a Tender 043057-2026, Digital Health coverage.
| Item | Detail |
|---|---|
| Reference | SBS10523 — Healthcare AI Solutions Open Framework |
| Buyer | NHS Shared Business Services Ltd, on behalf of NHS SBS Approved Organisations (NHS bodies, social care, plus authorised public/private bodies; UK + Guernsey + IoM + Jersey) |
| Value | £750m ex VAT / £900m incl. VAT |
| Term | 26 May 2027 – 11 May 2035 (7y 11m 17d) |
| Submission deadline | 21 July 2026, 12:00 noon BST (official, Bidders’ Instructions V3) |
| Enquiry deadline | 12 Jun 2026 (passed) |
| Portal | SAP Ariba — RFx 1110014540 |
| Award decision | 19 March 2027 |
| Procedure | Open Framework (Procurement Act 2023) — Votee + Beever are “Key Parties” |
| Call-off mechanisms | Direct award (no competition) OR competitive selection / mini-tender |
| Max suppliers | Unlimited per lot |
| NCIS levy | 0.90% on all sales |
| SME suitability | Flagged across all 8 lots |
| Lots | 8 lots (Lot 8 = automatic for suppliers winning ≥2 lots) |
2.1 The 8 Lots (per official notice)
| # | Lot | Award criteria | BritiAI/Votee/Beever fit | Recommendation |
|---|---|---|---|---|
| 1 | Radiology & Diagnostic Imaging | TBC | Low — no UKCA-marked imaging model | Skip |
| 2 | Pathological Diagnosis & Early Detection & Preventative Healthcare | TBC | Low — no validated pathology model | Skip |
| 3 | Virtual & Robotic Health | TBC | Medium — WardFlow Copilot via Votee agentic | Stretch (Wed 24 Jun decision) |
| 4 | Predictive Analytics | TBC | Medium — Beever Atlas + MAGIC LLM | Stretch (Wed 24 Jun decision) |
| 5 | Research, Innovation & Development | Quality 90% / SV 10% / Commercial PF | High — Beever ACL 2025 + HKUST/Vector/Stanford | Primary bid |
| 6 | Operational Efficiency | Quality 90% / SV 10% / Commercial PF | Strongest fit — V-Note, Copilot, Atlas, FOI Triage | Primary bid — quality lead |
| 7 | Advisory & Specialised Support | Quality 70% / SV 10% / Commercial 20% (scored) | High — Fractional CAIO + Implementation Pods; services-only lot — no medical-device claim required | Primary bid — pricing matters |
| 8 | Combined Solutions | Quality (auto) | Automatic for ≥2 lot winners | No separate bid — flows from 5+6+7 |
2.1a Confirmed evaluation weightings (Award Questionnaire V1.1)
Common to every lot: QL1 Scalable & capable 12.5% · QL2 Key challenges 10% · QL3 Data identification 13.75% (Quality subtotal 36.25%) + SV1 “Skills for Growth” 10% (the entire Social Value score on every lot).
- Lot 6 (Quality 90% / SV 10% / Commercial Pass-Fail): L6.1 Operational efficiency improvement 13.25 · L6.2 Technical support 7.25 · L6.3 Quality standards 13.25 · L6.4 Oversight 10 · L6.5 Interoperability 10. Heaviest levers: QL3 (13.75) and L6.1/L6.3 (13.25 each).
- Lot 7 (Quality 70% / SV 10% / Commercial 20% scored): L7.1 Service delivery · L7.2 Change management · L7.3 Clinical/operational involvement in design · L7.4 Risk approach · L7.5 Capabilities — 6.75 each. The 20% Commercial is scored, so the rate card genuinely matters here.
- Lot 5 (Quality 90% / SV 10% / Commercial Pass-Fail): L5.1 Collaboration 10 · L5.2 Safety 10 · L5.3 Product development & research design 13.75 · L5.4 User involvement 10 · L5.5 Interdisciplinary teams 10.
2.2 Mandatory Standards & Evidence
Standard NHS framework expectations (assume mandatory until ITT pack confirms; treat any “missing” as immediate red risk):
- DTAC v2 (Digital Technology Assessment Criteria) — self-assessment + evidence
- DCB0129 — Clinical Risk Management for manufacturers (Clinical Safety Officer required)
- DCB0160 — Awareness for deployment-side (mainly trust-side, but referenced)
- ISO 27001 — Information Security Management
- Cyber Essentials Plus — mandatory for NHS data handling
- NHS Data Security and Protection Toolkit (DSPT) — annual submission
- MHRA registration / UKCA marking — only if claims fall under medical device regulation (a major reason to avoid Lots 1, 2)
- ICO registration & DPIA capability
- NHS England AI Knowledge Repository alignment
- Financial standing: typically 2× contract value annual turnover OR parent company guarantee OR consortium aggregation
- Insurance (confirmed, Appendix 1): Employer’s Liability £5m / Public Liability £5m / Professional Indemnity £2m
3. Capability Audit — Who Brings What
3.1 BritiAI (britiai.com)
- 7 productised AI capabilities: Social Listening, Copilot, ASR, Image Detection, Document Processing, Report Generation/BI, Data Processor
- Horizontal cross-industry positioning (finance, healthcare, retail)
- UK-incorporated commercial vehicle — natural prime contractor
- Gaps: no published NHS case studies, no DTAC, no DCB0129, no ISO 27001, no Cyber Essentials Plus shown, no healthcare-specific marketing
3.2 Votee AI (votee.ai)
- MAGIC — enterprise LLM training on proprietary data (“vibe-driven” / 95% agentic training pipeline)
- V-Note — meeting transcription + summarisation, on-prem deployment supported
- Votee AI Studio — autonomous agent platform
- Agentic RAG chatbot with voice
- XProp — vision + OCR for document extraction
- Multimodal Theorem + LLM Evaluation Benchmark (defensible “we measure our models” story)
- HQ Hong Kong, Toronto AI lab, ex-Deloitte AI Institute Chief Scientist (Sep 2025)
- Gap: non-UK HQ — needs UK data residency story + DPA structure
3.3 Beever AI (beever.ai) — Votee’s research lab
- MAGIC CLI — agent system for low-resource & industrial LLM training
- TheoremExplainAgent — multimodal explanation videos (ACL 2025)
- HKCanto-Eval — benchmark (ACL 2025 CoNLL)
- Beever Atlas — open-source neural memory layer (chat → structured org knowledge via semantic + graph storage)
- Collabs: HKUST, Vector Institute, Stanford, Waterloo, HKU
- Strategic value: the evidence base. Peer-reviewed work, benchmarks, and IP we can cite in any Quality response. Critical for Lot 5.
3.4 Combined “stack” we sell into the NHS
| NHS pain point | Our stack answer |
|---|---|
| Clinician documentation burden | V-Note (ambient capture) + BritiAI ASR + Copilot summary |
| Discharge letter / referral backlog | XProp + Document Processing + Report Generation |
| Trust-level knowledge fragmentation | Beever Atlas (neural memory layer) |
| Tailored NHS LLM that doesn’t leak data | MAGIC training pipeline on sovereign UK infra |
| Service-line ops & demand forecasting | BritiAI Report Generation + Predictive (Lot 4 stretch) |
| AI implementation capacity gap | BritiAI Advisory + Fractional COO model (Lot 7) |
| End-to-end programmes | Integrated combined offer (Lot 8) |
4. Bid Structure — Recommendation
Recommended structure: BritiAI as Prime, Votee + Beever as Named Subcontractors under a binding Teaming Agreement.
Why this beats the alternatives
| Option | Pros | Cons | Verdict |
|---|---|---|---|
| A. BritiAI prime, Votee+Beever subcontractors (recommended) | Single UK contracting entity; clean DPA chain; fastest to execute; consortium financials can still be aggregated via parent guarantee | BritiAI carries prime liability | CHOOSE |
| B. Formal consortium / SPV | Shared liability optics; pooled balance sheet | Requires new legal entity; 4 weeks insufficient to incorporate and credit-rate | Reject (time) |
| C. Votee prime, BritiAI UK presence | Strongest tech narrative | HK-headquartered prime triggers data sovereignty objections; harder DSPT | Reject (risk) |
| D. BritiAI reseller of Votee only | Simple | Excludes Beever research story (which is our Lot 5 differentiator) | Reject (weakens score) |
Teaming Agreement — must contain
- IP licence-in from Votee/Beever to BritiAI for the framework term + 2 years tail
- Data processing flow-down clauses (UK GDPR Art 28)
- Exclusivity within the framework lots we name them on
- Pricing schedule and revenue share (proposed: 60% BritiAI / 30% Votee / 10% Beever on Lot 6 ops; 30/50/20 on Lot 5 research; advisory 80/15/5)
- Joint Clinical Safety Officer arrangement
- Step-in rights if any party becomes unable to deliver
- Heads of Terms signed before Monday’s session ideally, full agreement signed by 9 June
5. Strategic Positioning — The Narrative
Tagline candidates (pick one Monday):
- “Sovereign AI for the NHS — built in Britain, evidenced in the lab.”
- “From ambient capture to operational intelligence — the agentic NHS, one stack.”
- “The AI workforce the NHS already has — just not yet activated.”
Three pillars to anchor every section of the response:
- Sovereign by design — UK data residency, UK-hosted inference, transparent model training (MAGIC), open evaluation benchmarks (Beever). Direct answer to NHS England’s AI assurance posture.
- Clinically safe without overclaiming — we deliberately bid non-medical-device lots first. DCB0129 + DTAC v2 + DSPT, no MHRA shortcut. Trust boards can deploy without device-class scrutiny.
- Research-grade evidence — Beever’s ACL-published work + benchmarks gives every claim a citation. Few competitors will have this.
Differentiators to hammer:
- Ambient documentation that runs on-prem (V-Note) — not a US cloud round-trip
- Neural memory layer (Beever Atlas) — institutional knowledge stays inside the trust
- Fractional COO model in Lot 7 — embedded operating partner, not just consultancy days
- Open benchmarks — we publish how our models score, competitors don’t
6. Bold Solution Concepts (for the bid narrative)
Per your steer to be bold and creative, the proposal will lead with named solutions, not capability lists. Each is engineered to be (a) deployable in a single trust within 90 days; (b) measurable against a published KPI; (c) re-skinnable across the framework call-off market.
- “Scribe-On-Site” — On-prem ambient clinical documentation (V-Note + BritiAI ASR + Copilot). KPI: ≥60 min/clinician/day reclaimed. Lot 6.
- “WardFlow Copilot” — Agentic discharge orchestration: drafts letters, books TTOs, flags blockers. Lot 6 (+ Lot 3 stretch).
- “Atlas for Trusts” — Beever Atlas deployed as trust-wide neural memory layer (policies, SOPs, M&M notes, clinical guidelines). KPI: ≥40% reduction in time-to-find. Lot 6.
- “MAGIC-NHS” — Sovereign LLM training pipeline letting trusts fine-tune on their own data without it leaving the boundary. Lot 5 + Lot 7.
- “Trial Concierge” — Agentic recruitment + protocol Q&A for clinical trials. Lot 5.
- “AI Readiness Programme” — 12-week embedded engagement: DTAC gap analysis, use-case prioritisation, executive coaching, board reporting. Lot 7.
- “Integrated Digital Front Door” — Combined offer wrapping 1+3+6 with a single trust SRO. Lot 8 flagship.
- “FOI & Complaints Auto-Triage” — Document processing + classification (low-risk, high-volume entry point). Lot 6.
7. Win Themes (used verbatim throughout the response)
- Time-to-value: ≤90 days from call-off to first measurable KPI
- No data egress: UK-sovereign inference, on-prem option for every solution
- Evidenced, not asserted: every model claim is benchmarked (Beever)
- Embedded, not delivered: advisory is an operating partnership, not slide decks
- Buy-once, deploy-many: framework-priced reusable building blocks
8. Commercial Strategy
Pricing posture: mid-market premium. Not the cheapest (we lose the quality framing), not the most expensive (NHS price scrutiny). Target ~10–15% below incumbent enterprise vendors on equivalent scope.
Pricing model offered in submission:
- Day-rate card for Lot 7 advisory (Junior £550 / Senior £950 / Principal £1,450 / Fractional Exec £2,200)
- Per-user/month for Lot 6 productised solutions
- Outcome-share option for Lot 8 (capped) — bold differentiator
- Volume discounts at 5/10/25 trust thresholds
Margin model assumes: Votee SaaS COGS ~25%, BritiAI delivery ~45% gross, blended ~52% gross margin on Lot 6 SaaS; advisory at 65%+.
Revenue scenario (post-award, conservative):
- Year 1 (2027–28): 3 trust call-offs × £350k = £1.05m
- Year 2: 8 trust call-offs avg £450k = £3.6m
- Year 3: 18 trust call-offs avg £550k = £9.9m + 1 Lot 8 programme £2m = ~£12m
- Cumulative 3-year contracted run rate: £16–18m
9. Risks (full revised RAID in document 02)
Top 5 by exposure under the 3-week plan:
| # | Risk | Mitigation |
|---|---|---|
| R1 | ISO 27001 / CE+ / DSPT not certified before submission | Low risk — confirmed by official Q&A. A bidder passes if it holds or has formally commenced the required accreditations (or equivalents); evidence is only checked at due-diligence (Apr–May 2027). CE+ or ISO 27001 is accepted as equivalent. BritiAI submits with all three in progress and committed, backed by Votee/Beever ISO 9001 (held) as partner QMS evidence and a dated CEO-signed roadmap. |
| R2 | DCB0129 Clinical Safety Officer unappointed by Fri 20 Jun | CSO identified — contracting in process. Finalise engagement letter by Fri 20 Jun. Confirm registration body + number + training certificate. |
| R3 | Lot 5 evidence insufficient | Wed 24 Jun go/no-go gate; keep Lots 6 + 7 clean rather than submit a weaker 3-lot pack |
| R4 | Votee HK HQ → data sovereignty challenge | UK-only deployment topology (already done per Status Report); UK sub-processors only; DPA flow-down in Teaming Agreement |
| R5 | Social Value Plan under-scored (10% of every lot) | Upgrade from Partial to fully scored response by Fri 26 Jun; PPN 06/20 alignment with measurable commitments |
10. Decisions Needed THIS WEEK (official deadline 21 Jul; early internal submit-by 3 Jul)
- ✅/❌ GO confirmed (CEO Status Report indicates progress underway)
- ✅/❌ Lot scope: 5 + 6 + 7 primary, 3 + 4 stretch (decision Wed 24 Jun). Lot 8 automatic
- ✅/❌ Bid structure: BritiAI prime, Votee + Beever named subs via Teaming Agreement
- ✅/❌ Budget envelope: ~£65k cash / ~£90k total — restored to original 4-week envelope minus internal time already spent
- ✅/❌ CSO engagement — standard rate, signed by Fri 20 Jun
- ✅/❌ Cadence: Tue + Thu 09:30 standups; CEO Wed + Fri 16:00 check-in
- ✅/❌ Pricing posture (mid-market premium; Lot 7 rate card sharpened for 20% commercial weight)
- ✅/❌ Heads of Terms with Votee + Beever signed by Fri 19 Jun; full Teaming Agreement Fri 26 Jun
- ✅/❌ NHS reference letter campaign — revived under three-week window; target 1 warm letter by Fri 27 Jun
- ✅/❌ Social Value Plan — upgrade priority (10% of every lot; PPN 06/20 alignment)
11. Annex — Target Trust List for Reference Letter Campaign + Post-Submission Warm-Up
The official 21 Jul deadline gives a comfortable pre-submission window for a single warm NHS reference letter. Outreach starts this week; target 1 signed quotable letter by Fri 27 Jun. Same trust list then doubles as post-submission market warm-up through Q3/Q4 toward direct-award conversations for go-live Mar 2027.
Prioritised for: AI-curious CIO/CCIO, active digital strategy, existing relationship potential.
- University Hospitals Birmingham NHS FT — PIONEER AI hub
- Great Ormond Street Hospital — DRIVE unit, AI receptive
- Imperial College Healthcare NHS Trust — iCARE, strong research
- Manchester University NHS FT — digital pathology + ambient interest
- NHS Lothian — Scottish anchor (also CivTech relationships)
- Chelsea & Westminster — known for fast pilot adoption
- Moorfields Eye Hospital — research-led, AI-friendly
- Royal Free London — historic AI partnerships
- Leeds Teaching Hospitals — Yorkshire & Humber digital lead
- Northumbria Healthcare — operational efficiency culture
Pre-submission approach (15 Jun – 4 Jul): warm intro asking for a quotable reference letter on existing or prior interaction, NOT a free pilot offer. Target 1 letter by Fri 27 Jun. Post-submission approach (8 Jul onwards): 1-page paid pilot offer framed as direct-award-ready under SBS10523, targeting one signed pilot by Sep 2026 to seed the framework’s first-year pipeline.