The research team framed generative AI as a form of support infrastructure, a conceptual move with practical consequences. Rather than treating a large language model as a standalone tool that must somehow suit every user equally, they treat it as an interactional support capability whose usefulness depends on policies governing tone, pacing, scope, and boundaries. This framing matters because real users are not stable, average consumers of information. Their cognitive load, emotional state, and situational demands fluctuate from hour to hour, and a system optimized for a hypothetical typical user will systematically fail those whose conditions vary most.

To make this problem empirically tractable, the researchers chose a critical case: career support for professionals with ADHD. Attention-deficit/hyperactivity disorder, understood within the broader neurodiversity spectrum of cognitive variation in attention, learning, and sensemaking, makes variability itself the defining feature of work life. Difficulties with task initiation, prioritization, and sustained attention become consequential precisely during execution, so the helpfulness of any support hinges on alignment with shifting cognitive, affective, and situational conditions. If an AI system can sustain usable, appropriately bounded support across that variability, the design logic behind it should generalize to many other contexts where user states fluctuate.

Methodologically, the study is notable for its echeloned design science research process, organized as a four-stage theory-development pipeline. In the first stage, the team specified the problem space, drawing on a naturalistic online subcommunity corpus to categorize support demands by type and orientation. The second stage derived empirically grounded design requirements from that evidence. The third stage operationalized those requirements as interaction policies embedded within a configured large language model. The fourth stage evaluated the resulting policies through both controlled experiments and ecological, situated evaluation, using four participant samples. The evidence base spanned two eight-week deployment rounds, with interviews and check-ins in the first round and structured written reflections in the second, allowing the researchers to trace how candidate policies behaved under naturalistic variability.

The culmination of this pipeline is MATRIX, a set of six interaction-level design principles: momentary regulation, ability amplification, trust and stability, resonant attunement, intentional flourishing, and experiential integration. Momentary regulation addresses the need to help users manage acute cognitive and emotional states in the present instant. Ability amplification shifts the system’s stance toward building on users’ existing strengths rather than compensating for deficits. Trust and stability concerns the consistency users need to rely on the system over time. Resonant attunement captures how responses must align with the user’s current condition rather than a generic template. Intentional flourishing orients support toward the user’s longer-term goals, and experiential integration ensures that support fits coherently into the flow of lived work rather than adding friction.

Technically, the deployed architecture is a layered, policy-governed pipeline rather than a monolithic prompt. System-level invariants establish stable constraints across every interaction, including a nonjudgmental tone, diagnostic restraint, preserved user autonomy, and bounded support. Above that, a provisional routing layer classifies observable textual cues about the type and orientation of support demanded, activating primary and compatible secondary policy bundles. A structured response schema then organizes each reply into bounded acknowledgment, actionable guidance, and optional reflection. Finally, boundary and calibration rules regulate epistemic stance, scope, and proportionality, defining conditions under which the system should defer or refer the user elsewhere. Together, these layers deliver context-sensitive support within consistent governance constraints.

The evaluation strategy deliberately combined controlled and ecological approaches, reflecting a long-standing tension in research on human-technology interaction: laboratory experiments offer internal validity, while situated use reveals whether designs survive contact with real life. In the ecological rounds, participants chose their own work scenarios and used the system over eight weeks, then described which interactional features they valued, which they questioned, and why those qualities mattered in neurodivergent work practices. Cross-round coding compared interview evidence from the first deployment with written reflections from the second, allowing the team to test the stability of the policies and refine their boundaries before abstracting them into the final principles.

The theoretical payoff is a mid-range design theory that relocates the locus of inclusive design. Instead of asking whether a model is smart enough, the theory asks how support is calibrated and bounded as user conditions change. Calibration here means maintaining proportionality among three things: the user’s need, the situational demand, and the system’s warrant—what it actually knows and can responsibly claim. Boundedness means the system knows when guidance should be structured and paced, and when it should acknowledge its limits. The researchers are explicit about scope: the theory’s transferability is bounded to episodic, text-mediated support characterized by cognitive and affective variability, where guidance must be structured, paced, and delimited under shifting demand. It is not a universal prescription for all AI applications.

The broader implications reach into workplace inclusion, human resources, and the design of everyday AI assistants. Neurodivergent professionals have long faced support systems designed around neuronormative assumptions, and prior research has documented both the costs of ADHD across the lifespan and the shortage of effective workplace interventions. A design approach that treats fluctuating user conditions as the norm rather than the exception could reshape how organizations deploy AI for coaching, task management, and career development—not only for neurodivergent employees but for anyone whose attention and energy vary, which is to say everyone. The study’s emphasis on interaction policies also offers a practical template: organizations configuring large language models can specify invariants, routing logic, response schemas, and calibration rules as auditable design artifacts.

There are honest limits. The datasets, protected by institutional review board requirements, are not publicly available, though the authors note that de-identified data may be shared under appropriate safeguards when IRB protocols permit. The MATRIX labels themselves were applied retrospectively for expository consistency; the stage-by-stage pipeline distinguished deployed policies from principles finalized through analysis, a transparency move that guards against circularity. And the illustrative reference implementation published with the paper marks several components as prospective extensions rather than evaluated features. Even so, the study lands at a timely moment. As generative AI becomes ambient infrastructure in knowledge work, the question is shifting from what these models can do to how their help should be governed. Inclusive intelligence offers one answer: build the guardrails into the conversation itself, so that support remains usable, trustworthy, and appropriately scoped no matter how the user’s day is going.

Subject of Research: Design theory for generative AI systems that provide calibrated, inclusive support to neurodivergent professionals with fluctuating cognitive conditions

Article Title: Generative AI as Support Infrastructure: A Design Theory of Inclusive Intelligence

Article References: Abhari, K., Shirazi, H., Safaeipour, M., & Segal, R. (2026). Generative AI as Support Infrastructure: A Design Theory of Inclusive Intelligence. Information Systems Frontiers. https://doi.org/10.1007/s10796-026-10824-1

Image Credits: AI Generated

DOI: 10.1007/s10796-026-10824-1

Keywords: generative AI, inclusive intelligence, large language models, ADHD, neurodiversity, design science research, interaction policies, career support, human-AI interaction, support infrastructure, Information Systems Frontiers, MATRIX

The theoretical backbone of the study is Organizational Information Processing Theory, a framework first articulated in the 1970s by Jay Galbraith and later refined by Richard Daft and colleagues. The theory holds that organizations are, at their core, information-processing machines. When the uncertainty and ambiguity, or equivocality, of the information they face exceeds their processing capacity, performance suffers. Hospitals are perhaps the most extreme example of this mismatch. A single patient encounter can generate data from laboratory systems, imaging archives, electronic health records, insurance databases, and referral letters, each in different formats and each carrying its own uncertainties. Clinicians must reduce this flood of raw data into a diagnosis, a treatment plan, and a discharge decision, often under time pressure. The authors describe these mismatches as information contingencies: situations where the amount, clarity, or reliability of information does not match what decision-makers need.

To translate this old theory into the AI era, the researchers conceptualize HAIC as a third-order construct, a layered structure in which a single overarching capability emerges from the interaction of two second-order pillars: information processing and information decisiveness. This architectural choice matters methodologically. Rather than treating AI capability as a vague, monolithic asset, the model forces researchers and executives to measure distinct, testable components. Information processing captures how well an organization absorbs, integrates, and makes sense of the data flowing through its systems. Information decisiveness captures how confidently and coherently the organization converts that processed information into timely decisions and coordinated action. In the authors’ framing, an AI system that generates brilliant predictions but leaves clinicians paralyzed by ambiguity has failed on the second pillar, no matter how sophisticated its models are.

Within the information processing pillar, the study identifies two critical drivers: predictive accuracy and data interoperability. Predictive accuracy is the familiar face of AI, the statistical fidelity with which a model forecasts outcomes such as patient length of stay, readmission risk, or disease progression. Interoperability is the less glamorous but arguably more decisive factor. It refers to the ability of AI systems to exchange data seamlessly across platforms, departments, and even institutions. The authors argue that both must advance together. A highly accurate model fed by fragmented, incompatible data sources produces confident nonsense, while perfectly interoperable data flowing into weak models produces noise at scale. Only the combination augments an organization’s genuine capacity to process information, reducing both the uncertainty of not knowing enough and the equivocality of having too many ambiguous interpretations at once.

The information decisiveness pillar rests on two different supports: decision support and organizational collaboration. Decision support refers to AI systems designed to augment human judgment, presenting clinicians with ranked options, risk scores, and explanations rather than opaque outputs. Research on physician interaction with AI has shown that the way recommendations are framed shapes whether doctors adopt, adapt, or ignore them. Organizational collaboration, meanwhile, extends decisiveness beyond the individual clinician. Healthcare delivery is a team sport involving physicians, nurses, administrators, insurers, and external partners. The study finds that when AI capability is paired with structures that encourage cross-functional coordination, the organization becomes more decisive as a whole, converting shared information into unified action rather than siloed hesitation.

The payoff of this capability, according to the study, is concrete and operational. The authors establish a positive association between HAIC and timely, cost-efficient service delivery, which in turn reinforces overall operational effectiveness in healthcare. In practical terms, a hospital with strong HAIC is better positioned to shorten waiting times, allocate beds and staff more intelligently, and reduce the ancillary costs that accumulate when decisions stall. This finding echoes a growing body of operations management research showing that analytics capabilities in hospitals influence cost, quality, and patient satisfaction, but the new study’s contribution is to specify the mechanism: it is the reduction of information contingencies that unlocks the operational gains, not the mere presence of AI tools.

Methodologically, the study is notable for how it handles the messy reality of measuring an organizational capability. The authors ground their construct development in a content-analytic examination of the healthcare AI landscape, drawing on documented evidence from the sector, and then validate the higher-order structure using partial least squares structural equation modeling, a statistical technique well suited to complex, multidimensional constructs. The appendices of the published paper reveal the depth of this work: correlation matrices, cross-loadings to establish discriminant validity, and robustness checks including endogeneity assessments using Gaussian copula approaches, a technique increasingly favored in management research when randomized experiments are impossible. The authors also employed structural topic modeling to analyze textual data from the healthcare AI ecosystem, ensuring that their constructs reflected real-world configurations rather than armchair theorizing.

The practical implications reach hospital executives directly. The study offers AI-enabled configurations, essentially evidence-based recipes, for how to assemble the components of HAIC. An organization strong on predictive accuracy but weak on interoperability should prioritize data integration infrastructure before buying more sophisticated models. An organization with excellent analytics teams but poor cross-departmental collaboration should invest in governance and coordination mechanisms, because decisiveness will remain the bottleneck. The framework also suggests a diagnostic use: by measuring the four sub-dimensions, administrators can identify precisely where their information pipeline leaks value, whether at the point of data capture, in the analytical engine, or in the human decision that follows.

For scholars, the study extends Organizational Information Processing Theory in a direction that has waited decades for the right technology. Classic OIPT research treated information processing capacity as a property of organizational structure, such as formalization, decentralization, and the richness of communication media. The new work shows that AI systems can serve as a fundamentally new information processing mechanism, one that operates at machine speed and scale, but only when embedded in complementary organizational arrangements. This reframing connects the study to the resource-based view of the firm, in which capabilities, not technologies, are the sources of sustained advantage. An AI model can be copied; the organizational capability to exploit it cannot.

The timing of this research is significant. Studies of AI implementation across United States hospitals published in 2026 reveal a landscape of uneven adoption, with many institutions deploying tools without the surrounding capabilities to extract value. Meanwhile, healthcare costs continue to climb globally, and universal health coverage remains out of reach for billions of people, according to World Bank monitoring. Against this backdrop, the message of Bizalwan, Kumar, and Bose is quietly radical: the future of AI in medicine will not be decided by the cleverness of algorithms alone, but by whether healthcare organizations can build the twin disciplines of processing information decisively and deciding on processed information. Hospitals that master both may finally convert the long-promised AI revolution into something patients can feel in waiting rooms and wallets alike.

Subject of Research: Healthcare artificial intelligence capability and its role in reducing information contingencies to improve decision-making and operational effectiveness in hospitals

Article Title: Healthcare Artificial Intelligence Capability: Translating Information Contingencies into Decision Effectiveness and Operational Impact

Article References: Bizalwan, D. S., Kumar, R., & Bose, I. (2026). Healthcare Artificial Intelligence Capability: Translating Information Contingencies into Decision Effectiveness and Operational Impact. Information Systems Frontiers. https://doi.org/10.1007/s10796-026-10820-5

Image Credits: AI Generated

DOI: 10.1007/s10796-026-10820-5

Keywords: healthcare AI, artificial intelligence capability, Organizational Information Processing Theory, information processing, information decisiveness, predictive accuracy, data interoperability, decision support, operational effectiveness, hospital management, health informatics, Information Systems Frontiers