Before a functional, intelligent, and sustainability-oriented CA can be developed, a solid foundation is required. This foundation is being built through a systematic literature review – a scientific process that ensures the conversational agent is developed based on proven knowledge.

The review follows the established method by Brendel et al. (2020), consisting of six well-defined steps:

Step 1: Preparation

In the initial phase, the objective of the review was defined:

To identify digital nudges that have proven effective in the scientific literature for encouraging sustainable behavior through digital assistants.

At the same time, working definitions for the core terms were established:

1. Nudging: A strategy for behavioral change through modifications in decision-making environments.
2. Conversational Agent (CA): A dialogue-capable, AI-based system that interacts with users in natural language and provides guidance or options for action.

Step 2: Define Scope

In the next step, the scope of the review was defined. This included:

• Publications from recognized academic journals and databases
• Focus on studies addressing digital nudging
• Inclusion of research on CA interactions and behavior change
• Relevance to sustainability and corporate applications

Step 3: Literature Search

The actual literature search was conducted using four well-known academic databases:

• IEEE Xplore
• AIS eLibrary
• Web of Science
• Scopus

This resulted in a total of 1711 potentially relevant studies. After removing duplicates, 1181 unique entries remained.

These were then screened based on titles, abstracts, and keywords. The outcome: promising studies to be included in the next steps of the analysis.

Step 4: Analysis – Coding & Categorization

The next step involves a detailed full-text analysis of the 76 studies. Each study is examined for the presence of digital nudges, which are then systematically coded.

This means that each type of nudge is described and assigned to an appropriate category. These categories form a structured overview that will later feed into the development of the conversational agent.

Examples of such categories may include:

• Reminders & Alerts (e.g. maintenance prompts)
• Default Options (e.g. pre-selected sustainable settings)
• Transparency & Feedback (e.g. visualizations of resource consumption)
• Personalization (e.g. contextualized, data-driven recommendations)

Step 5: Synthesis

In this phase, the findings from the analysis are synthesized. The goal is to identify overarching patterns and best practices:

• Which nudging mechanisms are most effective?
• What combinations of dialogue and visualization support sustainable choices?
• What role do trust, clarity, or personalization play?

These findings will form the blueprint for the later prototype development of the digital assistant.

Step 6: Discussion

Finally, the results are reflected upon in light of the project goals and current academic discourse:

• To what extent can existing nudging strategies be transferred to asset management in organizations?
• What gaps still exist in research?
• What ethical implications arise from the use of digital nudges?
• What Does This Mean for ReCAIcle?

The systematic review ensures that ReCAIcle is not developed in a vacuum. Rather, the project is grounded in evidence-based insights into how digital systems can influence human behavior in meaningful and responsible ways.

1. In practical business applications, this could look like:
2. The CA politely reminding employees not to replace devices unnecessarily
3. Recommending appropriate repair options before malfunctions occur
4. Analyzing usage data to suggest resource-saving actions

Importantly, the users always remain in control – the assistant merely nudges them in the right direction through intelligent, data-informed suggestions.

From Theory to Practice

The literature review is just the beginning. Based on its findings, the project team will develop and test prototypes, and evaluate them in real-world settings with actual users. The ultimate goal: a tool that helps organizations practice sustainability intelligently – with a digital sparring partner by their side.

ReCAIcle thus combines technological innovation with psychological insight, behavioral strategy, and ecological responsibility. The systematic literature review is the scientific compass guiding the project – and a strong example of how interdisciplinary research can drive real-world change.