Boundary-Objects Account

• Boundary-Objects Account is an approach that defines technological artifacts as mediators facilitating flexible yet disciplined interactions across diverse stakeholder groups.
• It employs mixed-method evidence, such as surveys and interviews, to demonstrate how chatbots balance privacy, continuity, and regulatory demands.
• The framework provides actionable insights for designing culturally tailored, adaptive digital mental health tools that address competing stakeholder needs.

Boundary-Objects Account

A boundary-objects account specifies how technological artifacts bridge multiple stakeholder groups, enabling shared, flexible, yet disciplined interaction across domains of practice. In contemporary design-tensions frameworks, boundary objects serve as mediating entities that can be interpreted and appropriated differently by adolescents, families, peers, and mental-health services, while nonetheless supporting continuous and consistent system-level outcomes. This approach is central in the assessment of culturally grounded mental-health chatbots for underrepresented populations, particularly Indian adolescents, as elucidated in recent mixed-methods research that explicitly operationalizes the boundary-object lens within a design-tensions framework (Sehgal et al., 11 Nov 2025).

1. Conceptual Foundations and Boundary Objects in Design Tensions

Boundary objects are defined as artifacts that are adaptable to different users’ needs and constraints, yet maintain enough coherence to serve as a persistent referent for collaboration. In the domain of adolescent mental-health chatbot design, the boundary-object lens identifies how chatbots can simultaneously mediate between adolescents’ private psychological needs, family expectations, social stigma, and institutional support requirements. Chatbots themselves become boundary objects—entities that are at once conversation partners, information proxies, and hand-off mechanisms to professional services, with affordances driven by domain-specific tensions (Sehgal et al., 11 Nov 2025).

2. Mixed-Methods Evidence: Elicitation of Design Tensions via Boundary Objects

A large-scale survey (n=362) and qualitative interviews (n=14) conducted among Indian adolescents demonstrate the practical consequences of treating chatbots as boundary objects. Adolescents emphasize privacy, conversation continuity, localized content, and emotional support. These requirements correspond to, and are often in tension with, regulatory constraints (e.g., DPDP 2023 privacy legislation), infrastructural affordances, and cultural practices. The chatbot, as boundary object, is instrumented to selectively surface, hide, and mediate these requirements, allowing its function and interpretation to shift according to stakeholder engagement.

3. Formalization of Tension Pairs Mediated by Boundary Objects

The Design-Tensions Framework operationalizes seven paired dimensions that chatbots must balance, each pair reflecting a boundary-object role:

Tension Pair	Stakeholder Demand A	Stakeholder Demand B
Privacy ↔ Memory	Anonymity, non-retention	Conversation continuity, selective history
Direct Advice ↔ Reflection	Actionable, concrete next steps	Open-ended, reflective prompting
Mixed-Initiative ↔ Control	Guided, scaffolded interactions	Freeform, self-directed interaction
Local Curation ↔ Link-Dump	Region-specific, summarized assistance	General resources, external links
Text-First ↔ Voice-Optional	Quiet, written mode preference	Occasional spoken interaction
Persona Fit ↔ Tone Discipline	Personalized “friend”/“mentor”	Restrained “expert”/depersonalized mode
Concise ↔ Sufficient Detail	Brief, succinct replies	Expanded, more informative depth

Each boundary-object (chatbot instance) must be configurable or adaptive across these axes, preserving the artifact’s shared infrastructure while permitting locally meaningful variations.

4. Boundary-Object Artifacts: Implementation Rubric and Interaction Patterns

Designers operationalize the boundary-object account by implementing the tensions as concrete interaction affordances, codified in an artifact-level rubric (as in Table 5 of (Sehgal et al., 11 Nov 2025)). For example, privacy is maintained by default via zero-retention modes and optional transcript exports; memory is enabled selectively via “Save this conversation?” toggles. Guidance and user control are balanced with skippable flows and free-text escapes, allowing boundary spanning between novice and experienced adolescent users. Persona fit and tone discipline are realized via user-selectable conversational styles, minimizing anthropomorphic ambiguity and permitting non-deceptive, configurable engagement. Progressive disclosure of information supports concise yet expandably detailed interactions.

5. Mediation Across Stakeholder Groups and Institutional Contexts

As boundary objects, chatbots enable navigation of regulatory, social, and infrastructural divides. Adolescents and families interact with the system as a confidential support space, while clinicians, counselors, and institutional partners (e.g., Tele-MANAS) access logs, hand-off features, or crisis pathways as dictated by context. The artifact’s interpretation flexibly adjusts—for instance, a “friend” persona for peer-like engagement, an “expert” mode for crisis hand-off—while backend support preserves secure, auditable records in alignment with privacy statutes (Sehgal et al., 11 Nov 2025).

6. Application Guidance: Evaluative and Generative Use of the Boundary-Objects Account

Researchers and designers utilize the boundary-object account in both evaluative and generative contexts. Evaluation proceeds via auditing existing chatbot artifacts along the tension pairs; systems are measured for explicit privacy controls, mode-switching capabilities, regional localization, and personalized interaction affordances. Generative design is conducted by actively surfacing each tension during co-design workshops, prompting stakeholders—especially adolescents—to articulate concrete preferences regarding how the boundary object (chatbot) should respond to competing demands (e.g., “How should the chatbot balance anonymous support against ongoing care continuity?”). Boundary objects, in this framework, serve as modular, scalable mediators, supporting adaptation to other LMIC populations and contexts with comparable stakeholder tensions.

7. Theoretical and Practical Significance

The boundary-objects account foregrounds the role of technological artifacts as mediators of complex, multidimensional trade-offs, especially in contexts of high cultural and systemic variability. By making tensions explicit, the framework facilitates systematic balancing of privacy, autonomy, cultural fit, and institutional constraint, transforming the chatbot from a generic digital aid into a rigorously designed boundary object tailored to nuanced, situational needs of underrepresented adolescent populations (Sehgal et al., 11 Nov 2025).