Triage-and-Booking AI at a UK Telehealth Provider — 71% Demand Deflection, Same-Day Clinician Slots
ChatNext + Clinical Pathway Engine delivering NHS-compatible triage-and-booking — clinically safe, demand-routed and culturally appropriate.
The challenge
The client — a UK telehealth provider delivering services to a mix of NHS and private-pay patient populations — was facing a structural demand-supply mismatch. The provider's clinician network (GPs and nurse practitioners delivering telehealth consultations) was constrained, while patient demand was growing faster than the network. Average wait-time for a clinician slot was 7 days for non-urgent appointments, with urgent cases being deflected to the NHS-111 service or to the NHS A&E system in a pattern the provider's clinical leadership considered both inefficient and clinically suboptimal.
The traditional approach — adding clinicians — was constrained by the structural shortage of UK GPs. The provider's clinical leadership had concluded that the answer was structural: a meaningful fraction of the demand reaching the clinician network did not actually require clinician judgement, and could be appropriately handled through self-care guidance, signposting to community pharmacy, signposting to alternative NHS services, or appropriate clinician handoff with the right context.
The constraints were clinical-safety-led. Any triage-and-booking AI had to operate within clinically-validated pathways — the provider's clinical leadership would not deploy a system that made clinical judgements outside evidence-based guidance. NHS Information Governance Toolkit compliance applied. The patient population spanned a wide range of digital literacy and language preferences. And the provider's CCIO had been explicit that the AI must not be perceived by patients as a gatekeeper denying them clinical care.
The approach
MindMap deployed a triage-and-booking platform composed of ChatNext (Cn) as the conversational triage interface, Clinical Pathway Engine (Cp) as the clinically-validated pathway logic, Patient Scheduler (Ps) as the booking-and-routing layer, and Multi-Channel Agent (Mh) for the WhatsApp, web and app channel mix.
Phase one was the pathway-library build. The provider's clinical leadership selected approximately 180 clinical pathways covering the high-volume conditions the service was seeing (minor illness, dermatological complaints, sexual health, mental-health screening, women's health, paediatric minor illness, men's health and a long tail). Each pathway was based on the relevant NICE Clinical Knowledge Summary or NHS clinical-pathway guidance, with the clinical leadership signing off the pathway logic.
Phase two was the conversational triage build. The triage conversation walks the patient through the clinically-validated pathway questions for their presenting complaint, with the conversation adapting to the patient's responses and the conversation tone calibrated to be appropriately empathetic without being clinically over-reassuring. The conversation language is plain-English with culturally-appropriate phrasing and accessibility support.
Phase three was the routing logic. Based on the pathway outcome, the patient is routed to one of: self-care guidance (with explicit advice on when to escalate if the condition worsens), community pharmacy (where the pathway indicates a pharmacy-managed condition), alternative NHS service (where the pathway indicates a different service provides better care), same-day clinician appointment (where the pathway indicates clinician judgement is required), urgent clinician appointment (where the pathway indicates rapid clinician contact is required), or 999 / A&E (where the pathway indicates a clinical emergency).
Phase four was the clinician-handoff context. For patients routed to a clinician appointment, the triage conversation produces a structured clinical summary that the clinician receives at the start of the appointment — the presenting complaint, the relevant history, the pathway findings, the system's confidence on the routing decision. This eliminates much of the clinician's history-taking time at the start of the appointment.
The pre-built building blocks
Rather than commission a ground-up build, the engagement leaned on MindMap's pre-built accelerator library — production-tested components that compress what would otherwise be a six-to-nine-month build into weeks.
ChatNext
Conversational triage interface in plain-English UK patient tone
Clinical Pathway Engine
NICE-aligned clinically-validated pathway logic
Patient Scheduler
Booking-and-routing across clinician and alternative-service destinations
Multi-Channel Agent
Web, app and WhatsApp channel unification
The architecture
The platform runs on the provider's UK-region Azure tenant with full NHS Information Governance Toolkit compliance. All patient interaction processing happens in the UK; no patient data is sent to any non-UK API.
ChatNext's conversational layer uses a fine-tuned Llama 3.1 70B variant trained on the provider's historical triage-and-consultation corpus (with appropriate de-identification). The model is constrained to the pathway logic — it cannot invent clinical pathways or generate clinical advice outside the validated pathway library — and is calibrated for the empathetic, plain-English tone the provider's clinical leadership has defined.
Clinical Pathway Engine encodes the 180 pathways in a structured format that the platform evaluates against the patient's conversation responses. Each pathway has explicit decision-points, evidence citations, and escalation triggers. The pathway library is version-controlled and clinically-reviewed on a defined cadence; pathway updates flow into production through a clinical-governance approval workflow.
Patient Scheduler handles the booking-and-routing across the multiple destination channels (clinician calendar, community-pharmacy referral, alternative-NHS-service signposting, urgent-care routing). The clinician-calendar integration uses the provider's existing scheduling platform; the alternative-NHS routing uses the relevant NHS service-finder APIs.
The audit trail captures every triage interaction, every pathway evaluation, every routing decision, and every clinician interaction outcome. The provider's clinical-governance team has direct access to the audit trail for clinical-safety review; the Care Quality Commission has been granted appropriate inspection access under the provider's regulatory framework.
The numbers behind the story
Approximately 71% of patient demand is now handled without consuming a clinician slot — either through self-care guidance, community-pharmacy signposting, alternative-NHS-service routing, or other appropriate pathway outcomes. The remaining 29% reaches a clinician slot, with same-day availability now the standard rather than the exception.
Clinical-safety outcomes have been validated through extensive clinical-governance review. The provider's clinical leadership has reviewed the platform's pathway-outcome distribution against the comparable manual-triage baseline and concluded that the platform's clinical safety is equivalent to or better than the manual baseline, with the additional benefit of consistent pathway-adherence that human triage cannot guarantee.
Patient-experience outcomes have improved meaningfully. The provider's patient-experience research shows that patients value the speed of the triage process, the explicit explanation of why they are being routed to a specific destination, and the certainty of same-day clinician availability when clinician care is indicated. The previous experience of being placed on a waiting list with uncertain wait-time was a major dissatisfaction driver that the new approach has eliminated.
Clinician experience has improved as well. The clinicians receive a structured clinical summary at the start of each appointment, eliminating the history-taking time at appointment start. The clinicians' working theory is that they are seeing a higher proportion of the cases that genuinely require their judgement, rather than spending appointment time on cases that could have been resolved through self-care guidance.
An unexpected outcome: the platform's pathway-outcome data has become a source of population-health insight for the provider. The provider's clinical leadership uses the aggregated pathway-outcome patterns to identify emerging public-health patterns (seasonal infections, mental-health demand surges, regional variations in specific conditions) faster than the previous appointment-based data would have surfaced them.
“We had a structural demand-supply problem that adding clinicians alone could not solve. MindMap delivered a clinically-safe triage-and-booking platform that deflects seventy-one per cent of demand from clinician slots and ensures same-day availability for the cases that need clinician care. Our clinical-governance team validated the safety profile and our patients' experience has improved meaningfully.”— Chief Clinical Information Officer· UK Telehealth Provider
Why MindMap was chosen
The provider had evaluated three triage-automation vendors. The leading global vendor's platform was strong on the conversational UX but its clinical-pathway library was US-centric and did not align with NICE and NHS clinical guidance. The two UK-headquartered vendors had the NHS-pathway alignment but lacked the conversational-UX depth.
MindMap's accelerator-composition approach — bringing ChatNext, Clinical Pathway Engine, Patient Scheduler and Multi-Channel Agent together with the NHS-IG-Toolkit-compliant UK deployment and the deep NICE-pathway integration — was the structural differentiator.
Our embedded UK clinical expertise on the delivery team (two former NHS GPs and a former NHS-Digital clinical-safety specialist) was the third factor. The provider's CCIO felt that the team understood the clinical-safety and clinical-governance realities of UK triage, not just the conversational AI technology.
Related deployments
Prior Auth Acceleration
Prior Auth Accelerator + DocGenie automated 70% of payer prior auth submissions end-to-end, reducing turnaround from 3 days to 4 hours.
Medical Records Processing
Medical Records Parser processing 14,000 patient documents per day across nine hospitals, lifting coding accuracy from 87% to 99.2%.
Clinical-Trial Document Workflow
DocuMage + Clinical Trial Matcher accelerated study-start document cycle 5x — letting trial sites enrol patients weeks sooner than the prior baseline.
Want an outcome like this?
Start with a 2-week AI Readiness Sprint. We deliver a prioritised use-case backlog and business case grounded in what's actually buildable with our accelerator library.