Scoping Review Identifies Core Requirements for AI Tools in Healthcare Dialogue

Healthcare institutions worldwide are integrating artificial intelligence applications into daily operations, particularly for managing communications with patients, families, and other requesters. A newly published scoping review examines what it takes to deploy these tools effectively in interpersonal exchanges. The analysis, titled “Requirements for an effective application of artificial intelligence in interpersonal communication in healthcare: A scoping review,” appears in the Zeitschrift für Evidenz, Fortbildung und Qualität im Gesundheitswesen and is available online as of 24 June 2026.

Authors Teuvo Hänig and Thomas Petzold screened 1,301 studies published between May 2020 and May 2025, ultimately including 49 that met rigorous criteria. Their work maps organizational, design, and individual-level factors that support trustworthy, ethical, and user-centered AI communication systems in clinical and administrative settings.

Why Communication Processes Matter in Digital Health Transformation

Every healthcare facility handles repeated, structured interactions with people seeking information, appointments, advice, or support. These exchanges range from appointment scheduling and symptom triage to post-discharge follow-up and administrative queries. When designed well, AI systems can handle routine elements, freeing staff for complex cases while maintaining empathy and accuracy.

The review highlights that successful deployment depends on more than technical capability. It requires deliberate attention to ethical frameworks, legal compliance, intuitive interfaces, and staff and patient competencies. Without these elements, even sophisticated tools risk eroding trust or introducing new inefficiencies.

Methodology: Applying the PCC Framework to AI Communication Literature

Hänig and Petzold followed established scoping-review guidance, using the Population-Concept-Context model to organize findings. They searched multiple databases and applied clear inclusion and exclusion rules focused on direct interactions between requesters and AI applications. The resulting 49 studies originated primarily from Europe (39 percent) and Asia (31 percent), with smaller contributions from North America, South America, Australia, and the Middle East.

This geographic distribution reflects both the rapid adoption of digital tools in certain regions and the growing international interest in evidence-based guidance for AI in sensitive health contexts.

Organizational Requirements: Ethics, Law, and Institutional Integration

A major cluster of findings centers on how healthcare organizations embed AI tools within existing governance structures. Key requirements include clear policies for data protection, transparent decision-making about when AI may interact directly with requesters, and alignment with national and international regulations such as the EU AI Act and data-protection directives.

Institutions must also define accountability lines: who is responsible when an AI system provides incomplete or misleading information? The review stresses that ethical and legal embedding cannot be an afterthought; it shapes every subsequent design and training decision.

Photo by Abdulai Sayni on Unsplash

Design Requirements: Usability, Transparency, and Performance

Effective AI communication tools need interfaces that feel natural and reduce cognitive load. Studies included in the review emphasize intuitive navigation, clear language tailored to diverse literacy levels, and visible indicators that users are interacting with an automated system rather than a human.

Performance expectations include rapid, accurate responses and graceful handoff to human staff when queries exceed the system’s scope. Visual and textual presentation of results must remain understandable, avoiding jargon that could confuse patients or caregivers.

Individual-Level Requirements: Data Literacy and Competency Development

Beyond systems and policies, the review identifies competencies required of both staff and requesters. Healthcare professionals need training to oversee AI outputs, recognize limitations, and intervene appropriately. Patients and families benefit from basic digital health literacy so they can interpret AI-generated information and know when to seek clarification.

Handling sensitive personal data responsibly emerges as a cross-cutting theme. Clear consent processes, easy access to data-correction mechanisms, and transparent explanations of how information is used all support sustained acceptance of these tools.

Evidence Gaps and Practical Implications for Healthcare Delivery

Despite the breadth of literature examined, the authors note an absence of direct evidence measuring how AI communication tools affect requester outcomes such as satisfaction, adherence, or health literacy. This gap underscores the need for future outcome-focused studies.

At the same time, broader adoption of well-designed systems could help address structural challenges, including staff shortages and unequal access to information in underserved regions. The review suggests that requirements identified here provide a practical checklist for pilot programs and scaling decisions.

Relevance for Academic Research and Professional Education

Health-services researchers, bioethicists, and informatics scholars will find fertile ground in the identified gaps. Universities offering programs in digital health, nursing informatics, or medical ethics can incorporate the review’s framework into curricula, preparing graduates to design, evaluate, and govern these technologies.

Postdoctoral and faculty positions in health AI governance are likely to grow as institutions translate these requirements into operational policies. The scoping review supplies a shared vocabulary and evidence base that can accelerate collaborative projects across disciplines and borders.

Readers interested in related career pathways can explore current openings in research and faculty roles focused on digital transformation in healthcare.

Photo by The 77 Human Needs System on Unsplash

Future Outlook: Toward Context-Sensitive, Human-Centered AI Communication

The authors conclude that sustainable success depends on holistic consideration of requirements during conception, development, and implementation. Context always matters: a tool effective in one hospital may need adaptation for another with different patient demographics, regulatory environments, or staffing models.

Ongoing dialogue among clinicians, patients, technologists, ethicists, and policymakers will be essential. The scoping review offers a timely foundation for that conversation and for evidence-informed decision-making as AI becomes a standard feature of healthcare communication worldwide.

Access the full publication here: ScienceDirect abstract and options. The work is credited to Teuvo Hänig and Thomas Petzold.