Evidence & measurement

Measure the gap.
Not just the speed.

No single number can capture human readiness. Combine experience, behaviour, workforce evidence, outcomes and governance quality.

What success looks like

People understand what is happening, can recover when it goes wrong, and leave the interaction with their confidence and control intact.

01 / Signal library

Seven evidence areas

Use multiple signals together. A high completion rate can still hide confusion, over-trust or assisted workarounds.

Human · 01

Understanding

Can people explain what the system is doing and what they should do next?

  • Comprehension testing
  • Explanation usefulness
  • Support questions
Human · 02

Confidence & trust

Does the journey increase confidence—or transfer it uncritically to automation?

  • User feedback
  • Repeat use
  • Complaint themes
Service · 03

Navigability

Can people complete, pause, recover and return under real conditions?

  • Abandonment
  • Time on step
  • Error recovery
Service · 04

Support demand

Where is the service creating friction, uncertainty or repeat effort?

  • Contact rate
  • Escalation volume
  • Assisted completion
Human · 05

Risk & vulnerability

Does the service recognise heightened pressure and respond proportionately?

  • Fraud interventions
  • Coercion signals
  • Complaint severity
Organisation · 06

Workforce readiness

Can staff use, explain and challenge AI responsibly?

  • Training depth
  • Challenge confidence
  • Quality review
Organisation · 07

Governance quality

Is oversight meaningful, accountable and evidence-led?

  • Audit trails
  • Fairness testing
  • Issue resolution

02 / Maturity model

From reactive
to adaptive

Maturity is the ability to sense, interpret and respond to widening gaps—not simply the amount of AI deployed.

1

Reactive

Human adaptation is considered after issues appear.

2

Controlled

Basic governance exists, but behavioural insight is uneven.

3

Supported

Support, accessibility and explanations are designed in.

4

Integrated

The framework shapes design, risk and operations.

5

Adaptive

Customer and workforce evidence continuously improves AI.

03 / Behavioural signals

Read with care

Observed behaviour is a prompt to investigate the service and context—not a diagnosis of the person.

Observed signalCould indicateSupportive response
Repeated pausesLoad, uncertainty or fear of errorSimplify, save progress, make help visible
Rapid completionTime pressure, stress or over-trustAdd review moments at high-risk points
Channel switchingTrust, access or complexity pressurePreserve choice and learn from the friction
Journey abandonmentOverload, lost confidence or fearOffer recovery and a simpler route back

Move from evidence to action

Review a live journey.

Use the interactive HARMONIC checklist to surface gaps and leave with a focused action plan.

Open the toolkit