Q5 Health Insurer Operating Model Demonstrator

Active members

184,260

+2.8% YoY

Claims cost

£48.6m

+7.4% YoY

Avg. resolution time

18.7 days

+3.2 days

Customer outcome score

82 / 100

-4 pts

Enterprise Signals

Signal	Evidence	Implication
Claims Musculoskeletal pathway	+21% duration, +14% cost	Investigate provider delay and evidence requests
Service Repeat contact	17.8% of callers contact again in 7 days	Likely process and information failure
Risk Vulnerable cohort	1,482 members show combined vulnerability indicators	Review intervention and communication model
Opportunity Prevention	£5.2m historic claims linked to potentially earlier intervention	Candidate prevention proposition

Claims Cost Trend

Illustrative monthly indexed cost

Decision Queue

Decision 01

Claims triage bottleneck
Model a controlled move of simple claims into AI-assisted triage.

Decision 02

Prevention pilot
Test earlier intervention for a high-cost musculoskeletal cohort.

Decision 03

Span of control
Review management layers in claims and service operations.

Target Operating Model

SenseLive signals across customers, claims, operations and regulation.

UnderstandCustomer, case and enterprise intelligence.

PreventEarlier interventions to reduce harm and avoidable cost.

ProtectRisk intelligence and AI-supported underwriting.

ResolveCase orchestration focused on outcomes.

LearnEvery decision and outcome improves the system.

Customer Passport

Member 004781

Cover	Comprehensive health protection
Tenure	14 years
Open claims	1
Risk status	Medium
Vulnerability	Temporary mobility limitation
Next action	Clinical triage review within 24 hours

AI Case Brief

Customer has an active musculoskeletal claim. Evidence suggests two avoidable hand-offs and a 6-day provider delay. Recommend clinical triage, consolidated evidence request and proactive customer update.

Human authority required
Clinical, coverage and adverse customer decisions remain outside autonomous execution in this demonstrator.

Interaction Timeline

Date	Event	Observed issue	Suggested action
03 Sep	Claim notified	None	Standard intake
05 Sep	Evidence requested	Duplicate request likely	Consolidate evidence
09 Sep	Provider referral	Queue delay	Escalate provider SLA
14 Sep	Customer call	Repeat contact	Proactive update

Open claims

6,842

Simple claims

41%

Automation candidate

High-complexity

12%

Human-led

Preventable delay

23%

Primary redesign target

Claims Friction Map

Stage	Avg. time	Primary friction	AI-enabled intervention
Notification	0.8 days	Manual classification	Automated intake and evidence extraction
Triage	3.4 days	Queue and hand-offs	Risk-based routing
Evidence	5.9 days	Incomplete / duplicate requests	Evidence completeness engine
Provider	6.1 days	External delay	SLA prediction and escalation
Decision	2.5 days	Policy interpretation	Grounded decision support

Scenario: Reduce Manual Claims Handling

Simple claims moved to AI-assisted handling: 30%

Management layers removed: 1

Estimated capacity released	38 FTE
Indicative annual operating value	£2.4m
Resolution time effect	-2.8 days
Control risk	Medium

Accountability Test

Automation should not simply remove roles. It should redefine where judgement, exception handling and accountability sit.

AI	Classification, evidence assembly, routing, drafting, monitoring
Human	Adverse decisions, clinical judgement, exceptions, vulnerability, appeals
Executive	Risk appetite, delegated authority, performance and control design

Future Organisation Logic

Current orientation	Future capability	Design consequence
Product	Customer & proposition	Organise around need and lifetime value
Underwriting	Risk intelligence	Move towards continuous risk understanding
Claims	Resolution & recovery	Measure outcomes, not queue movement
Customer service	Customer orchestration	Single journey across channels
Compliance	Continuous assurance	Controls embedded in execution
IT / Data	Digital operating platform	Shared data, orchestration and evidence layer

Prevention Service

MSK Early Intervention

Target members showing early indicators before claim severity increases.

Navigation Service

Care Pathway Navigator

Help customers reach the right provider and reduce avoidable hand-offs.

Employer Proposition

Workforce Risk Insight

Anonymised organisational risk and prevention service for employer clients.

Opportunity Engine

Opportunity	Evidence	Value hypothesis	Next test
MSK prevention	£5.2m historic cost linked to delayed intervention	Lower severity and faster recovery	500-member pilot
Claims navigation	17.8% repeat contact	Lower service demand, higher confidence	AI-guided journey trial
Employer risk insight	Existing corporate relationships	New recurring service revenue	Co-design with 3 clients

Authority Ladder

Level	AI authority	Example	Control
1. Observe	No action	Detect claims delay pattern	Evidence only
2. Recommend	Advice	Propose case route	Human decision
3. Approve	Draft action	Prepare customer response	Human approval
4. Execute	Bounded action	Send low-risk status update	Policy + audit trail
5. Verify	Post-action check	Outcome and control verification	Continuous assurance

Required Evidence

Every material AI-supported action should retain the evidence, model or rule, recommendation, authority, decision, accountable person and outcome.

Design Principle

Use AI to increase organisational capacity and decision quality. Do not delegate accountability to the technology.

Explore

What is happening?

Interrogate claims, customers, service, risk, workforce, cost and propositions.

Explain

Why is it happening?

Trace signals to processes, cohorts, operating constraints and evidence.

Design

What could we change?

Generate operating-model, process, service and organisational options.

Discover

What else could we do?

Identify new services, products, interventions, partnerships and revenue models.

Question The Organisation

The system should answer in a management format: evidence, interpretation, options, risks, dependencies, next experiment and decision owner.

Question type	Example
Diagnostic	Why has claims duration increased?
Comparative	Which customer cohorts create most avoidable cost?
Counterfactual	What happens if we automate 40% of simple claims?
Design	Design a lower-cost claims operating model.
Discovery	What else could we offer these customers?
Challenge	What assumptions in our current model are weakest?

Management Response Pattern

1. Signal	What the system has observed.
2. Evidence	Source data and confidence.
3. Interpretation	Likely causes and uncertainties.
4. Options	Alternative interventions.
5. Impact	Customer, cost, risk and workforce effects.
6. Next Test	Smallest useful experiment.
7. Authority	Who decides and who is accountable.

What Else Could We Do?

Prevention

Move before the claim.
Identify cohorts where early intervention could reduce severity, duration or avoidable absence.

New Propositions

Monetise expertise.
Create employer risk insight, care-navigation, prevention and specialist advisory services.

Operating Model

Redesign around decisions.
Remove hand-offs, collapse unnecessary layers and create cross-functional outcome teams.

Provider Network

Manage the ecosystem.
Use outcomes, timeliness and cost to steer referrals and intervene before service failure.

Customer Experience

Move from contact centre to orchestration.
Give members one intelligent journey across service, claim, intervention and recovery.

Knowledge Asset

Turn 100 years into usable intelligence.
Capture policy, claims, clinical, operational and expert knowledge into an evidence-backed organisational memory.

From Question To Action

AskNatural-language management question.

RetrieveFind relevant governed evidence.

ReasonIdentify patterns, assumptions and uncertainty.

OptionGenerate alternative interventions.

SimulateEstimate effects and dependencies.

DecideHuman authority, action and verification.

Executive Command

Active members

Claims cost

Avg. resolution time

Customer outcome score

Enterprise Signals

Claims Cost Trend

Decision Queue

Decision 01

Decision 02

Decision 03

Target Operating Model

Customer Passport

AI Case Brief

Interaction Timeline

Open claims

Simple claims

High-complexity

Preventable delay

Claims Friction Map

Scenario: Reduce Manual Claims Handling

Accountability Test

Future Organisation Logic

Prevention Service

Navigation Service

Employer Proposition

Opportunity Engine

Authority Ladder

Required Evidence

Design Principle

Explore

Explain

Design

Discover

Question The Organisation

Management Response Pattern

What Else Could We Do?

Prevention

New Propositions

Operating Model

Provider Network

Customer Experience

Knowledge Asset

From Question To Action