Unit 3: EMPATHY - Deep Human Understanding with AI Support

I. Orientation: Empathy as the First Mode of Design Thinking

Empathy is the opening phase of the five-mode design thinking model (Empathize → Define → Ideate → Prototype → Test) popularised by the Stanford d.school and IDEO. It is the discipline of setting aside the designer's assumptions and grounding decisions in the lived experience of real users.

• Definition: Empathy is the ability to understand and share the feelings, needs, motivations and contexts of the people you design for — cognitive (understanding their reasoning) and affective (feeling their emotion).
• Human-centred premise: solutions must fit people, not the reverse; the user is the expert on their own experience.
• Divergent stance: empathy widens the problem space before Define narrows it — you gather more than you will use.
• Latent needs: users often cannot articulate what they want ("faster horse" vs. the car); observation reveals needs interviews miss.
• Output of the phase: rich qualitative data — quotes, observations, stories — later distilled into insights and a problem statement.

II. Empathy in the Design Process

A. It's Importance

Empathy prevents teams from solving the wrong problem efficiently.

• Reduces bias: replaces the designer's mental model with evidence from users, avoiding the "I am the user" fallacy.
• De-risks investment: early understanding is cheap; a misjudged need discovered after launch is expensive.
• Reveals latent needs: surfaces unspoken frustrations and workarounds users treat as normal.
• Aligns the team: shared user stories give cross-functional teams a common reference point for decisions.

B. Roots in Ethnography

Design empathy borrows its methods from anthropological ethnography — the study of people in their natural setting.

• Fieldwork, not lab: data is collected in situ — the kitchen, factory floor, hospital ward — where behaviour actually occurs.
• Emic perspective: meaning is understood from the participant's viewpoint, in their own words and categories.
• Thick description: Clifford Geertz's term for recording behaviour with its full context and meaning, not just the surface action.
• Participant observation: the researcher watches and partly joins the activity to grasp tacit, unspoken routines.

C. Empathetic Immersion

Immersion means deliberately experiencing the user's world to feel constraints first-hand.

• Shadowing: following a user through a task or day to see real sequences and pain points.
• Analogous immersion: entering a comparable extreme situation (e.g. designers wearing gloves to mimic arthritic hands).
• Extreme users: studying edge cases (heavy users, non-users) whose exaggerated needs expose issues invisible in average users.
• Contextual inquiry: interviewing while the user performs the task, so answers stay concrete rather than idealised.

III. AI Support for Empathy Work

A. Using AI for Thematic Analysis of Interviews

AI accelerates the coding of large volumes of qualitative interview data into recurring themes.

• Transcription & NLP: speech-to-text converts recordings, then natural-language processing tags entities, sentiment and topics.
• Automated coding: clustering and topic-modelling (e.g. grouping semantically similar statements) surface themes across dozens of transcripts fast.
• Sentiment tagging: flags emotional peaks — frustration, delight — that mark high-value moments.
• Workflow:

1. Transcribe audio → text
2. Segment into statements / utterances
3. AI clusters statements into candidate themes
4. Researcher reviews, merges, renames, discards
5. Extract representative quotes per theme

• Human-in-the-loop: the designer validates and interprets; AI proposes, the human decides.

B. AI-assisted Empathy: Strengths and Risks

AI extends reach and speed but cannot itself feel; strengths and risks must be weighed together.

1. Strengths:
   • Scale: processes hundreds of interviews or reviews that a team could never read manually.
   • Speed: compresses days of manual coding into minutes.
   • Pattern detection: finds cross-cutting themes and correlations a tired human eye misses.
   • Consistency: applies the same coding logic uniformly across all data.
2. Risks:
   • Loss of nuance: tone, sarcasm, cultural context and body language are flattened or lost.
   • Bias amplification: models trained on skewed data reproduce and scale existing prejudice.
   • False empathy: AI simulates understanding without genuine affective insight — automation bias makes teams over-trust it.
   • Privacy: interview data may contain sensitive personal information requiring consent and secure handling.

IV. Empathy Artifacts: Personas, Scenarios and Maps

A. Creating Personas & Scenarios

Personas and scenarios turn scattered research into shareable, memorable representations of users.

• Persona: a fictional but evidence-based archetype of a user segment — includes name, photo, demographics, goals, frustrations, behaviours.
• Grounding: every persona attribute traces back to real interview or observation data, never invention.
• Scenario: a narrative describing how a specific persona uses the product to achieve a goal in a realistic context.
• Use: keeps the abstract "user" concrete during ideation — "What would Priya need here?"

B. Customer Journey Mapping

A journey map visualises the end-to-end experience of a persona across time and touchpoints.

• Stages/phases: the horizontal axis — e.g. Awareness → Consideration → Purchase → Use → Support.
• Layers: for each stage record actions, thoughts, emotions (often an emotional curve), and touchpoints.
• Pain points & opportunities: low points on the emotional curve mark where design can add most value.
• Scope: covers the whole relationship, including moments outside the product itself.

C. Maps & Blueprints

Different maps expose different views; a service blueprint extends the journey map behind the scenes.

1. Journey / experience map: front-stage — what the customer sees, does and feels.
2. Service blueprint: adds back-stage — front-stage actions, backstage employee actions, and the line of visibility separating what the customer sees from internal support processes and systems.
   • Empathy map (see IV.D): a snapshot of one moment, not a timeline.
   • Choice rule: use a journey map for the experience over time, a blueprint to diagnose operational gaps causing pain points.

D. Developing Empathy Map & Journey Map

These are the two core synthesis artifacts built directly from field data.

• Empathy map — four quadrants around the user:

SAYS   |  THINKS
-------+--------
DOES   |  FEELS

• Says: verbatim quotes the user spoke aloud.
• Thinks: beliefs they hold but may not voice.
• Does: observed actions and behaviours.
• Feels: emotional state, often with pains and gains added.
  • Building the empathy map: cluster sticky notes from research into each quadrant; contradictions between Says and Does signal deep insight.
  • Building the journey map: plot the persona's stages, then populate actions, emotions and touchpoints from the same research, marking pain points for the Define phase.

V. Insighting

A. Importance of Insighting

Insighting is the act of moving from raw observations to a non-obvious understanding of why users behave as they do.

• Data ≠ insight: "40% abandon the cart" is a fact; "users abandon because hidden shipping costs feel like a betrayal" is an insight.
• Actionable: a good insight points toward a design opportunity, not just a description.
• Surprise test: the strongest insights are non-obvious and reframe how the team sees the problem.
• Tension-based: insights often sit on a contradiction between what users say and what they do.

B. Insighting through Interviews

Semi-structured interviews are the primary tool for eliciting the stories that yield insights.

• Open-ended questions: "Tell me about the last time you…" invites stories rather than yes/no answers.
• Five Whys: repeatedly asking "why?" drills past surface reasons to root motivations.
• Neutral probing: avoid leading questions ("Don't you hate…?"); let the user supply the frame.
• Listen for emotion & extremes: hesitations, strong words and stories mark where insight hides.
• Capture verbatim: exact quotes preserve meaning for later synthesis.

VI. Defining the Problem

A. Defining Problem Through Lens of Design Thinking

The Define mode converts empathy findings into a single, human-centred problem statement.

• Synthesis: cluster empathy-map and journey-map findings into themes, then into insights.
• Human-centred framing: the problem is stated in terms of a user's need, never a solution or feature.
• Right scope: narrow enough to act on, broad enough to allow many solutions.
• Bridge role: Define funnels the divergent empathy data into a focused brief for divergent ideation.

B. 'Point of View' Statement & 'How Might We' Statement

These two paired devices lock the problem definition and open it to ideation.

1. Point of View (POV) statement: a fixed, meaningful problem definition combining user, need and insight.

[User] needs [need] because [surprising insight].

• User: the specific persona, e.g. "A busy first-time parent".
• Need: a verb, an unmet goal, never a noun/solution — "needs to feel confident", not "needs an app".
• Insight: the because — the non-obvious reason drawn from research.
  2. How Might We (HMW) statement: reframes the POV as an open, optimistic question to launch ideation.

How might we [help user] [achieve need] [given insight]?

• How: presumes a solution exists.
• Might: invites many possibilities without commitment.
• We: signals collaborative ideation.
• Goldilocks scope: too broad ("HMW improve travel") gives no traction; too narrow ("HMW add a button") pre-decides the solution — pitch between the two.