## Objectives Klosterfrau wanted to understand what consumers know about and how they evaluate different OTC pharmaceutical ingredients. Beyond this contextual question, another objective was methodological: understanding the similarities and differences between AI-moderated interviews via chat and conventional online surveys. More precise, the goal was to determine whether the two methods produce comparable results and where each method might provide unique value. Ultimately, Klosterfrau wanted to improve ingredient-level decision-making by combining quantitative results with qualitative consumer context. ## Method & Approaches & Innovation In this project, 4 parties collaborated: Klosterfrau (industry partner), Q Agentur für Forschung (Agency), horizoom (Online Access Panel) and xelper (AI tool provider). We conducted a two-arm validation study with n=1,000 respondents in a conventional 15-minute online survey and n=1,000 respondents in a fully AI-moderated chat interview. Both arms used the same screener and identical set of ~40 OTC ingredients. The innovation of AI-moderated interviews lies in the combination of qualitative insights and quantiative reliability. The AI-moderator probes, asks for clarifications, and encourages participants to articulate their reasoning. This conversational approach produces rich, contextual opinions at a volume typically unachievable in traditional qualitative setups. Fieldwork runs 1–2 weeks and captures both deterministic measurement and rich narratives. Importantly, this large-scale comparison study represents a methodological innovation in its own right. It systematically benchmarks AI-moderated interviews and conventional surveys under identical experimental conditions. To the best of our knowledge, no prior research has conducted such a direct comparison — particularly at this scale and level of control. ## Results Full results will be available at GOR. Based on our hypotheses, we expect: - High comparability on core quantitative metrics (ingredient awareness, basic evaluations). - Higher efficiency of the conventional survey for deterministic items and structured ratings. - Strong added value from the AI-moderated interviews, generating richer narratives, decision rationales, misconceptions, and contextual associations that numbers cannot capture. The study will provide the first systematic, large-scale comparison of AI-moderated interviews versus conventional surveys. ## Impact For Klosterfrau, the project fundamentally enhances how decisions about products and active ingredients are made. The AI-moderated interviews reveal how consumers become aware of certain ingredients, what they associate with them, where misunderstandings occur, and how these factors shape purchase decisions. These insights directly inform ingredient communication, packaging and label claims, educational messaging, and the strategic prioritization of ingredients within the product portfolio. Beyond the immediate project outcomes, Klosterfrau and the broader research community benefit from the methodological validation itself. By empirically comparing AI-moderated interviews and conventional surveys, the study provides an evidence base for how AI-driven qualitative methods can be positioned within established research frameworks.

Objectives – Which business question did the client want answered?
Our client asked us to develop a strategy for the introduction of M365 Copilot. In addition to a quantitative employee survey of around 800 employees, additional employee interviews were to be conducted with the aim of determining user acceptance and preparing strategic action areas for the rollout to approximately 13,000 employees. For the generation, execution, and evaluation of these interviews, the client used the AI interview bot developed by us.

Method & Approaches & Innovation – How were the insights gathered?
The interview bot consists of a total of nine highly specialized AI agents, trained to conduct a real-time interview on the use of M365 Copilot. What makes this special is that the multi-agent model, like in a real conversation, can individually adapt to the mood, disposition, and needs of the interview partner and, following the DICE methodology (Descriptive, Idiographic, Clarifying, Explanatory), can ask suitable follow-up questions.

Results – What are striking and impactful insights?
With our solution, we succeeded in reducing interview costs from 400 euros to 0.5 euros. In addition, all information is broken down into small knowledge units using special logics and stored in optimized knowledge databases. With these technologies, we can utilize all semantic and thematic connections and make them available in an evaluation bot. By using this AI, market researchers can analyze information in the shortest possible time and ask follow-up questions at any desired level of detail.

Impact – How did the project move the needle for the client? What was done differently afterwards?
Following our project, the market research and innovation management departments became aware of us. In addition to further developing the interview bot, we are jointly planning to evolve the solution into a generic application for the generation, processing, and provision of corporate knowledge, which can be used in various areas of the company.

Objectives:
This study examines three approaches to traditional, use-case-based usability testing: AI-moderated interviews, human-moderated interviews, and unmoderated testing. The aim is not only to compare the quality of insights obtained but also to analyze user perception and participant behavior during interaction with the tools. An additional goal for the research team was to gain practical experience in setting up surveys, managing workflows, and exploring analysis capabilities.

Method & Approaches & Innovation:
In a controlled experiment with nine participants, identical tasks were carried out under all three test conditions. Collected metrics included objectively measurable dimensions such as word count in open-ended responses, user evaluations of the methods/tools (e.g., enjoyment, willingness to participate again, usability), as well as qualitative assessment of responses regarding depth, emotional engagement, and relevance.

Results:
Initial findings indicate that AI-based moderation offers significant advantages in scalability and cost but still shows usability shortcomings and limited applicability. It delivered high-quality responses and—contrary to our hypothesis—was broadly accepted, even among premium customers.

Impact:
The results contribute to the ongoing discussion about the maturity of AI-driven interview tools for UX research applications. Furthermore, they provide practical insights into their limitations and potential.

Objectives

Which business question did the client want to answer?

The German Women’s National Soccer Team (GWNST) needed reliable, budget-friendly insights to boost fan engagement and guide their marketing strategy leading up to UEFA Women’s EURO 2025. Their core question was:

How can we increase the popularity, emotional connection, and reach of women’s soccer using insights that go beyond traditional and AI-based research?

Method & Approaches & Innovation

How were the insights gathered?

The team pioneered a completely new methodology called Deep Implicit, designed to capture pure human intuition — something neither traditional research nor AI can reliably access.

The innovative approach included:

1. Guided audio-based alpha-state induction using techniques used in hypnosis and meditation and binaural beats to calm the rational mind and activate intuitive processing.
2. Projective questions asked during a relaxed state to elicit first, unfiltered System-1 thoughts.
3. Strict 30-second capture window to prevent rationalisation.
4. LLM-assisted multi-stage analysis:

• Filter for genuine System-1 responses
• Summarize intuitive statements
• Interpret deeper psychological meaning

1. Quantitative validation via a 500-respondent causal driver analysis using Supra Causal AI to test whether intuitive findings causally explain real-world fan excitement.

This is the first-ever methodology to measure intuitive qualitative insight in a standardized and automatized form, filling a major gap in both AI and market research.

Results

What are striking and impactful insights?

The Deep Implicit process uncovered a new strategic truth that traditional research and AI had missed:

1. Fans become supporters when they truly identify with the players as relatable humans — not as polished athletes.
2. They also need to identify with the team’s values — the club must “stand for something I support.”
3. Authenticity and emotional storytelling—not scripted content—are the most powerful drivers of fan excitement.
4. Quantitative validation showed:

• The new attribute “the club stands for something I support” was the #1 causal driver of fan excitement — surpassing all expert-designed items.
• Authenticity ranked among the top three drivers.

These insights fundamentally reframed how women’s soccer should communicate its identity.

Key takeaways:

• Without Deep Implicit, the Causal AI study would have missed to by far largest driver
• Without Deep Implicit, supporting LLMs would have come up with shallow inputs

The true power is in combining Human and Artificial Intelligence.

• Human intelligence is not only exert and domain knowledge, but the underleveraged intelligence of human is also called INTUITION
• Artificial intelligence is not just generative AI but causal AI, to explore ground truth in data.

Impact

How did the project move the needle for the client? What changed afterwards?

The insights were implemented immediately in the GWNST’s social media and communications strategy — shifting from scripted content to authentic, emotionally resonant, player-centered storytelling.

The effect was dramatic:

Instagram Performance (May → June)

After implementing the new strategy:

• Reach increased by +120% on average
  • Album posts: 296,878 → 542,928
  • Reels: 297,896 → 822,160
• Interactions nearly doubled
  • Albums: 9,908 → 19,206
  • Reels: 10,799 → 20,104

This represents a step-change in communication efficiency, achieved without additional budget — purely through better insight.

Beyond performance metrics, the project also:

• Provided the DFB with the first validated intuition-based insight tool
• Revealed a new strategic narrative for women’s soccer
• Demonstrated how AI and research can be complemented — not replaced — by intuitive intelligence

From Insights to Impact: A Life-Centric, AI-Driven Approach to Modern Brand Tracking

Modern market research increasingly faces the requirement to deliver not just insights but actionable, impact-oriented strategies. Our contribution addresses the central question of how brand tracking can be advanced to explain when and why consumers choose brands, how these decisions are embedded in both situational and general life contexts, and how such understanding can be translated into growth-oriented actions.

Methodologically, the approach builds on a life-centric brand tracking framework that captures brand usage within contextualized real-life situations and is guided by the concept of Category Entry Points (CEP) as defined by Byron Sharp. By applying a reversed questioning logic—asking “which brand fulfills which need in which situation?” rather than capturing image associations—we can model situational decision-making with greater precision. In addition, real user profiles are enriched using the AI system AIlon, which generates contextual information from consumers’ living environments and creates synthetic consumer profiles. Combined with social and behavioral science expertise, this results in a data-driven, context-sensitive model of brand choice.

The results show that situational needs are key drivers of brand decisions and that a life-centric approach significantly sharpens the understanding of these decision dynamics. AI-generated profiles deepen the contextual knowledge of consumers’ life situations and enable new activation points along the customer journey. The integration of tracking data, synthetic profiles, and AI-based context expansion increases the precision of target group modelling and makes consumers more effectively addressable.

For companies, this provides clear added value: instead of descriptive insights, they receive concrete communication and action strategies that are directly linked to situational consumer needs. Looking ahead, this approach positions brand tracking as a systematic, AI-supported steering instrument that identifies growth opportunities and aligns marketing activities with consumers’ real-life contexts.