What makes a smart artifact feel smart?
This research investigates how the emotional and cognitive dimensions of user experience shape people's perception of smartness in everyday AI-powered artifacts, arguing that perceived smartness is not a technical property, but a user experience outcome constructed through interaction, meaning-making, and identity.
ROLE
UX Researcher (PhD Study)
Methods
UX Research, Phenomenographic Study, Qualitative multi-method
FOCUS
Cognitive UX, Emotional UX
THE PROBLEM SPACE
"Smartness is not what an artifact does,
it's what a user feels."
Smart artifact domestication is transforming homes into hybrid physical-digital environments, reshaping users' behaviors, expectations, and mental models. Yet how individuals decide whether a artifact is "smart" has been largely overlooked in literature: prior work focused on technical specifications and usability metrics, not on the lived, emotional, cognitive experience of smartness
This research was designed to fill that gap, not to define smartness objectively, but to understand the full qualitative range of how people experience and attribute it.
Research Question 1
How do cognitive and emotional aspects of user experience influence the perceived smartness of smart artifacts?
Research Question 2
How do affordances and different interaction modes of smart artifacts affect user experience and perceived smartness?
Research Question 3
How do individuals give meaning to smart artifacts, their user experience, and their smartness in relation to interaction modes?
SELECTING THE
EXEMPLARY SMART ARTIFACTS
Before any fieldwork, I developed a Smartness Characteristics Definition Map, a synthesis of smartness definitions from prior literature, structured through thematic analysis into three characteristic groups: technical, performative, and definitive.
15 smart artifacts collected from Turkish e-commerce platforms were systematically evaluated and scored against this map. The three artifacts with the most distinct interaction modes, affordance profiles, and smartness profiles were selected.
Autonomous · mobile · physical
Robot vacuum cleaner
Operates independently of the user, takes over physical cleaning labour, and navigates domestic space: raising questions about autonomy, context-awareness, and trust when things go wrong.
Autonomous · mobile · physical
Smartwatch
Attached to the body at all times, it blurs the boundary between user and artifact: surfacing questions of identity, physical enhancement, and what it means to be permanently connected.
Autonomous · mobile · physical
Smart speaker
Interacted with entirely through voice, it demands a new grammar of use: raising issues of language competency, personality attribution, and the feeling of having an AI presence in the home.
Why these three?
Each artifact offers a distinct combination of
-
interaction modality (physical, wearable, voice)
-
participation type (active/passive)
-
sensory engagement (tactile, visual, auditory)
enabling the study to explore perceived smartness across qualitatively different UX conditions within the same research design.
RESEARCH
DESIGN
A Multi-phase Phenomenographic Inquiry
The methodology is grounded in phenomenographic research, chosen specifically for its capacity to surface the qualitative diversity of how different individuals experience the same phenomenon, rather than converging on a single universal account.
The method treats experience as simultaneously objective and subjective, and is structured around how participants conceptualize and give meaning to reality.
Three complementary research phases were sequenced to triangulate findings and progressively deepen analytical focus.
PHASE 1
PILOT STUDY
Validating the methodology
Two participants used all three selected artifacts over a condensed period. The pilot was designed not to generate findings, but to stress-test every research instrument: checking whether diary prompts surfaced emotionally rich moments, whether interview questions reached the depth of meaning-making required, and whether the mapping task produced analytically comparable outputs.
Key revisions included replacing a labor-intensive printed diary with WhatsApp for real-time capture, restructuring questions across multiple tools to reduce cognitive load, and simplifying the mapping task design.
PHASE 2
ONLINE RESEARCH STUDY
Mining unsolicited narratives
To complement the researcher-led case study, I collected and analyzed a large corpus of user-generated product reviews for all three device categories from major Turkish e-commerce platforms (Trendyol, Hepsiburada, n11 and Amazon).
This unsolicited data revealed how real users describe their smart artifact experiences when no researcher is present, offering a naturalistic triangulation layer.
Using Atlas.ti, I applied a structured three-round thematic analysis: first fragmenting narratives into meaningful units, then relating sub-themes to emotional and cognitive UX dimensions, and finally grouping them into conceptual categories.
This phase directly shaped the thematic coding framework used in the main case study.
PHASE 3
MAIN CASE STUDY
Longitudinal, multi-tool, naturalistic
Ten participants used all three artifacts over a one-week period in their actual home environments.
The study was structured in two phases:
01.
SHARING MOMENTS
In the first, participants were invited to contact the researcher when they encountered something remarkable: sharing moments, photos, and voice recordings via WhatsApp, then completing structured questionnaires in Google Forms.
02.
MULTI-TOOL ONLINE SESSIONS
In the second phase, a one-hour semi-structured interview session was conducted online, followed by a one-hour mapping study in Miro with a simultaneous think-aloud protocol. Interview questions were designed to surface narrative storytelling and meaning-making about smartness; the mapping task asked participants to explicitly link specific device functions to their perceived level of smartness while verbalizing their reasoning.
Participant selection required prior experience with robot vacuum cleaners and functional English fluency for smart speaker interaction, which lacked Turkish language support at the time of study.
ANALYSIS
RESULTS
The Anatomy of Perceived Smartness
All gathered data was analyzed through a three-round thematic analysis applied consistently across WhatsApp responses, questionnaire answers, interview transcripts, and mapping study recordings.
The rounds moved from:
narrative fragmentation → relational theming → conceptual grouping
Seven key concepts emerged, organized under emotional and cognitive UX dimensions. These are not taxonomic categories, they are experiential conditions: the situations and qualities that cause a device to register as smart (or not) in the mind of the users.
Emotional User Experience
Emotional UX focuses on the affective responses, feelings, and psychological bonds formed during interaction. The analysis investigates how smart artifacts trigger emotional changes, such as satisfaction, trust, or frustration, and how these experiences lead users to anthropomorphize or develop personal connections with technology.
By analyzing user narratives, this study identifies key concepts like emotional engagement, and intriguing existence that define the emotional dimension of smartness.
Naming tendency
1. Mr. Kaplan, 2. The Tasmanian Devil, 3. Katya.
Cognitive User Experience
Cognitive UX encompasses elements such as mental workload reduction, alignment with mental models, and the active interpretation of affordances.
The analysis shows how artifacts enhance a user's sense of competence and liberation by taking over complex tasks or providing gratifying experiences that exceed expectations.
The findings highlight how cognitive efficiency, perceived phenotype, and trust-building experiences contribute to the overall construction of perceived smartness from a user-centered perspective.
Competency enhancement
The need to intervene in errors.
Connectivity
Smartwatches acting as smartphone extensions.
Connectivity
Disconnection implies being "not smart".
DESIGN
IMPLICATIONS
The study demonstrates that perceived smartness is not a technical specification: it is a relational, experiential outcome. Users do not evaluate smartness by reading feature lists; they construct it through accumulated emotional and cognitive experience with the device over time.
This has two direct implications for practice: first, that the early phases of smart artifact design must include frameworks for evaluating emotional and cognitive UX alongside functional performance; and second, that academic frameworks for smartness must be grounded in user perspective, not engineering capability.
For researchers
This study contributes a conceptual framework with the seven experiential conditions that can inform future qualitative and quantitative research into smart artifact UX.
It also demonstrates the validity of phenomenographic methods for HCI research, and the utility of multi-tool longitudinal designs (WhatsApp diaries, think-aloud mapping, semi-structured interviews) for capturing the full complexity of naturalistic smart artifact experience.