This post explains what happened when researchers used strategy mapping software to run a participatory evaluation of COVID-19 crisis management and what teams doing AI-enabled qualitative research should learn. The primary keyword for this post is ai-enabled qualitative research and the audience is qualitative researchers and evaluation leads who run workshops, syntheses, and recommendation processes. The payoff: concrete, reproducible process steps, dates, participant counts, and software notes you can reuse in your next participatory evaluation.
Key Takeaways
According to PLOS ONE (published August 20, 2026), a strategy mapping workflow run in a collaborative browser tool produced ten participant-developed recommendations from four workshops with 15 healthcare professionals in April–June 2023 and was still perceived as impactful in follow-up interviews in spring 2024 (PLOS ONE).
- 15 participants took part in four strategy mapping workshops conducted on 3 April 2023, 6 April 2023, 24 April 2023, and 12 June 2023, according to PLOS ONE (August 20, 2026).
- The process resulted in 10 recommendations produced and validated by participants, according to PLOS ONE (August 20, 2026).
- Ten semi-structured follow-up interviews were conducted in spring 2024 to assess perceived use and value, according to PLOS ONE (August 20, 2026).
- The authors concluded that “strategy mapping methodology is a viable approach for conducting impactful participatory evaluations, ” (Björnqvist et al., 2026) as reported in PLOS ONE (August 20, 2026).
What happened and how the strategy mapping workflow worked
Answer: The study used strategy mapping software in four facilitated workshops to elicit risks, map causal links, and co-develop recommendations with stakeholders, according to PLOS ONE (August 20, 2026).
According to PLOS ONE (August 20, 2026), the evaluation targeted a regional communicable disease unit in Sweden and involved all 15 staff members in the unit across managers, nurses, and physicians.
According to PLOS ONE (August 20, 2026), the facilitator led a four-step process: blind individual idea generation, open gathering and clustering, causal linking with software analytics to find central risks, and participant rating of risks and recommendations.
According to PLOS ONE (August 20, 2026), the workflow used Strategyfinder (version 1.0.0) as the browser-based strategy mapping software and used the software’s centrality and loop-detection tools to select subsystems for recommendation development.
According to PLOS ONE (August 20, 2026), AI was used only for translation and figure recreation: the authors used M365 Copilot (based on GPT-5) to translate Swedish statements to English and to recreate visual maps, and the outputs were verified by the authors.
Findings Snapshot
| Date | Metric | Value | Implication |
|---|---|---|---|
| 3 Apr 2023 | Workshop 1: blind gather + causal mapping | 15 participants started; initial complete risk system created | Generated the baseline causal map used across workshops, according to PLOS ONE (August 20, 2026) |
| 6 Apr 2023 | Workshop 2: validation and subsystem creation | Five subsystems identified using software centrality measures | Prepared targeted recommendation development, according to PLOS ONE (August 20, 2026) |
| 24 Apr 2023 | Workshop 3: recommendations generation and participant ratings | Participant ratings of perceived effectiveness and feasibility guided selection | Produced candidate recommendations for validation, according to PLOS ONE (August 20, 2026) |
| 12 Jun 2023 | Workshop 4: in-person validation | Final validation and edits produced 10 final recommendations | Deliverable: ten participant-validated recommendations, according to PLOS ONE (August 20, 2026) |
| Spring 2024 (recruitment Feb–Apr 2024) | Follow-up interviews | 10 of 15 participants interviewed for ~1 hour each | Assessed perceived use and process value, reported in PLOS ONE (August 20, 2026) |
| 20 Aug 2026 | Publication | PLOS ONE article (Björnqvist et al., 2026) | Peer-reviewed case study with dataset archived at Zenodo, according to PLOS ONE (August 20, 2026) |
Implications for qualitative researchers and evaluators
Answer: Strategy mapping software can increase stakeholder ownership and produce actionable recommendations, but facilitators must balance participant authorship with recommendation clarity, according to PLOS ONE (August 20, 2026).
According to PLOS ONE (August 20, 2026), participatory mapping increased individual and collective reflection, and many interviewees described the method as “innovative and useful” when compared to prior evaluations.
According to PLOS ONE (August 20, 2026), drawbacks included a lengthy four-session cadence that some participants found slow, and recommendations that some interviewees called too general or dependent on other organizations.
According to PLOS ONE (August 20, 2026), practical choices to improve impact include: defining what counts as “use” before follow-ups, shortening the session count where feasible, and adding a facilitator-led editing step to increase the clarity and actionability of participant-written recommendations.
How Evidano helps AI-enabled qualitative research
What Evidano is
Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents.
According to the needs identified in the PLOS ONE case (Björnqvist et al., 2026), teams need reliable transcription, translation, thematic coding, rapid synthesis, and secure data controls to scale participatory evaluations.
Problem: messy transcription and translation → Solution: verified speech-to-text and translation
Answer: The PLOS ONE study used automatic transcription in Microsoft Word then manual correction, showing a workflow opportunity that Evidano can streamline.
According to PLOS ONE (August 20, 2026), interviews were transcribed with Microsoft Word’s speech-to-text and manually corrected; Evidano’s transcription features Speech-to-text with custom dictionaries and PII redaction reduce manual correction time and preserve traceability.
Problem: synthesizing workshop maps and interview codes → Solution: thematic + cross-segment analysis
Answer: The PLOS ONE case produced maps and later a directed content analysis; Evidano automates code extraction, frequency reporting, and cross-segment comparisons to accelerate that synthesis.
According to PLOS ONE (August 20, 2026), the authors coded interviews into predefined categories and summarized frequencies in Excel; Evidano automates thematic extraction, supports hierarchical codes and subcodes, and exports analytic tables suitable for validation and follow-up.
Problem: translating participant maps to clear recommendations → Solution: AI-assisted drafting + human-in-the-loop editing
Answer: Björnqvist et al. (PLOS ONE, August 20, 2026) recommended more facilitator involvement to improve recommendation clarity, which an AI-assisted workflow can enable while preserving ownership.
Evidano’s AI chat over your documents and analyses lets evaluators propose edited recommendation drafts, then track changes and send them back to participants for validation, matching the follow-up step proposed in the PLOS ONE discussion.
Problem: data security and reproducibility → Solution: encrypted, non-training storage and archival exports
Answer: The PLOS ONE dataset was stored with restricted access on Zenodo and shared under conditions that protect confidentiality, showing the need for secure handling.
Evidano provides enterprise controls and clear audit trails; see Data security for governance and export features that support controlled sharing similar to the Zenodo arrangement described in PLOS ONE (August 20, 2026).
Learn more about Evidano features
Answer: For teams planning to run participatory evaluations, Evidano’s feature set maps directly to the study’s needs.
See the product feature list at Evidano features and the speech and translation tools at Speech-to-text and Translation to prototype a shortened, AI-assisted participatory workflow inspired by the PLOS ONE case.
FAQ: ai-enabled qualitative research
How did researchers use software to turn discussion into recommendations?
Answer: They used Strategyfinder to capture blind and open gathers, cluster statements, draw causal links, and run centrality analytics, according to PLOS ONE (August 20, 2026).
The method combined individual blind idea-generation (to avoid early anchoring) with open sharing and facilitator clustering, then causal linking and participant ratings to prioritize recommendations, as reported in PLOS ONE (August 20, 2026).
Can AI be used to summarize or translate participant maps safely?
Answer: Yes, but with verification: the authors used M365 Copilot (GPT-5) for translation and map recreation and verified outputs against originals, according to PLOS ONE (August 20, 2026).
The PLOS ONE authors explicitly noted the AI outputs were checked by humans and were not used for analysis or interpretation, showing a human-in-the-loop approach is recommended.
What made the recommendations more likely to be used?
Answer: Participatory ownership and ratings increased perceived relevance, while clarity and dependencies reduced usability, according to PLOS ONE (August 20, 2026).
Interviewees told the authors that recommendations felt timely and helpful when they confirmed or structured ongoing work, but some were judged too vague or dependent on other organizations to be actionable, as reported in PLOS ONE (August 20, 2026).
How should evaluators define “use” when measuring impact?
Answer: Evaluators should define whether “use” means project initiation, plan incorporation, or practice change before follow-up, as recommended by the authors in PLOS ONE (August 20, 2026).
The PLOS ONE study found differing participant interpretations of “use, ” which complicated impact assessment in the follow-up interviews conducted in spring 2024.
Conclusion & Next Steps
Answer: The PLOS ONE case shows strategy mapping software can produce actionable, stakeholder-owned recommendations, but facilitators and AI-assisted editorial steps are needed to raise clarity and shorten timelines (Björnqvist et al., PLOS ONE, August 20, 2026).
If you run participatory evaluations, consider: (1) pre-defining what counts as “use” before follow-up interviews, (2) using AI-assisted transcription and translation with verification to save time, and (3) adding a facilitator editing pass followed by participant validation to improve recommendation actionability, as discussed in PLOS ONE (August 20, 2026).
Evidano can help you prototype this hybrid workflow through verified transcription, thematic and cross-segment analyses, and AI chat over your documents; see Evidano features to plan a pilot.
To try an AI-assisted qualitative workflow that maps directly to the study’s needs, Try Evidano for free.
Topics
- ai-enabled qualitative research
- strategy mapping software evaluation
- participatory evaluation crisis management
- qualitative analysis AI tools
Keep reading
- Commentary on NewsAI for Qualitative Analysis in Participatory EvaluationAI qualitative analysis participatory evaluation: Learn how AI-enabled qualitative analysis speeds synthesis of workshops and interviews from a PLOS ONE case study. Try Evidano.
- Commentary on NewsAI for Qualitative Analysis of Museum VisitorsAI turns observation notes into insights: qualitative analysis of museum visitors that finds themes, age segments, and exhibit recommendations for UX teams
- Commentary on NewsBoost Exhibit Design: AI Qualitative Analysis for MuseumsUse AI qualitative analysis for museums to convert observational studies into targeted exhibit changes. Learn from PLOS ONE findings and try Evidano to accelerate insight.