AI UX Case Study
Profile A – Find the right Free Travel application form
AI UX transformation at a glance
- Context: Public service guidance for identifying the correct Free Travel application form.
- Challenge: Reduce uncertainty and cognitive overload without altering eligibility rules.
- Approach: Layer structured, adaptive UX enhancements onto governance-verified content.
- Result: Clearer starting points. Lower decision anxiety. Faster route identification. Zero policy drift.
Baseline condition
The original topic was accurate, complete, and governance-verified.
However, users were required to simultaneously:
- Scan a dense comparison table
- Distinguish between multiple legitimate forms (FT1, FTM, FT U/70, FTNI1, FTNI2)
- Interpret cross-border distinctions (Republic of Ireland vs Northern Ireland SmartPass)
- Self-determine eligibility before acting
For many users, particularly older adults or those supporting someone else, the dominant emotional signal was:
“I hope I am choosing the correct form.”
Accuracy was strong.
Confidence was fragile.
The objective was not to change policy.
The objective was to reduce friction while preserving policy truth.
Enhancement Architecture
This case study demonstrates a structured three-layer AI augmentation model:
- Persona Mode Toggle — establishes context (emotional clarity)
- Quick Match Assistant — narrows decision space (decision clarity)
- Read Aloud Mode — reinforces comprehension and reassurance (comprehension clarity)
Each layer addresses a distinct friction point while preserving policy integrity.
Enhancement 1 — Persona Mode Toggle
Clarifying intent before filtering information
Problem signal
Users approach the same page with different mental models:
- Applying for themselves
- Supporting someone else
- Acting in a professional capacity
The baseline treated all users identically.
Design response
A persona context toggle was introduced at the top of the page:
- Users declare context before engaging eligibility content
- Tone and emphasis adapt accordingly
- No rules are changed
- No logic is altered
- No content is permanently hidden
All adaptation operates strictly in the presentation layer.
Impact
- Reduces emotional uncertainty
- Establishes immediate relevance
- Builds reassurance before decision-making begins
Enhancement 2 — Quick Match Assistant
Structured narrowing without rule distortion
Problem signal
The Free Travel scheme includes multiple valid routes:
- Standard age-based application (FT1)
- Medical route (FTM)
- Companion arrangements (FT U/70)
- Northern Ireland SmartPass routes (FTNI1 / FTNI2)
In the baseline model, users processed all routes simultaneously via a full comparison table.
Design response
A Quick Match Assistant was layered above the table:
- Structured triage prompts narrow likely pathways
- Decision cues are presented in plain language
- The full comparison table remains visible for transparency
The assistant does not replace official guidance.
It reduces cognitive load before detailed comparison begins.
Impact
- Lower scanning burden
- Reduced selection error
- Clearer cross-border differentiation
- Full transparency of official routes preserved
Enhancement 3 — Read Aloud Mode
Multimodal comprehension support
Problem signal
Dense policy tables and structured text can create friction for:
- Users with visual strain
- Users who process information more effectively through listening
- Users seeking reassurance via guided pacing
Reading is not the only valid comprehension pathway.
Design response
A read-aloud panel was integrated:
- Page content can be listened to in full
- Follow-along highlighting reinforces comprehension
- Playback controls allow speed adjustment
No duplication.
No policy rewriting.
Inclusive support layered onto verified content.
Impact
- Improved accessibility
- Increased confidence through guided interpretation
- Alternative cognitive pathways enabled without altering meaning
Governance Integrity
This transformation operates under strict structural constraints:
- Underlying eligibility rules remain unchanged
- All adaptive behaviour operates strictly within the presentation layer
- Track 1 governance authority is preserved
This is structured augmentation.
Not automated decision-making.
Outcomes
User outcomes
- Reduced entry-point anxiety
- Faster route identification
- Lower cognitive scanning load
- Clearer cross-border understanding
- Greater confidence in form selection
Structural outcomes
- Zero policy reinterpretation
- Full alignment with Department of Social Protection guidance
- Transparent adaptive logic
- Governance and UX operating as complementary tracks
What This Case Study Demonstrates
A disciplined model of AI-enabled UX in regulated environments:
- Adaptive interface behaviour layered onto verified content
- Friction reduction without rule distortion
- Transparent guidance rather than automation
- Multimodal accessibility integrated by design
- Governance-safe transformation suitable for public service contexts
The result is a public service page that is:
- More usable
- More reassuring
- More inclusive
- Fully policy-aligned
AI maturity here is not about replacing human judgement.
It is about designing adaptive clarity around trusted rules.