AI for Qualitative Analysis in Participatory Evaluation

Evidano is an AI-powered qualitative data analysis platform that helps researchers analyze interviews, open-ended surveys, and documents. According to the PLOS ONE case study (Björnqvist et al., 2026), using strategy mapping software in a participatory evaluation produced ten participant-developed recommendations and measurable process use. The following post explains how the PLOS ONE case (published 20 August 2026) ran four workshops in April–June 2023, followed up with ten interviews in spring 2024, and how AI-enabled qualitative research tools can reproduce and scale those steps for teams and evaluators.

Key Takeaways

According to the PLOS ONE case study, strategy mapping software used in a participatory evaluation produced ten participant-developed recommendations and generated identifiable use and learning among stakeholders.

15 healthcare professionals participated in a four-workshop series held on 3 April 2023, 6 April 2023, 24 April 2023, and 12 June 2023, according to PLOS ONE (Björnqvist et al., 2026).
The participatory process produced ten recommendations in April–June 2023 and the authors conducted ten follow-up interviews in spring 2024 to assess use, as reported in PLOS ONE on 20 August 2026.
Björnqvist et al. (PLOS ONE, 2026) report mixed views: many interviewees said recommendations were helpful and used to start projects, while some described recommendations as too vague or COVID-19 specific.
The study used an LLM (M365 Copilot, GPT-5) for translation and map recreation only, and the authors verified the model outputs against originals, as stated in the Methods section of PLOS ONE (2026).

What Happened and How the Evaluation Worked

What happened: Björnqvist et al. (PLOS ONE, 2026) applied a strategy mapping methodology via browser-based software to run four participatory workshops with a single unit for communicable disease and infection control, producing ten recommendations.

According to Björnqvist et al. (PLOS ONE, 2026), the workshops combined blind and open idea gathers, participant causal linking, subsystem extraction based on centrality, and participant ratings of risks and recommendations to prioritize actions.

According to Björnqvist et al. (PLOS ONE, 2026), the research team then conducted ten semi-structured follow-up interviews in spring 2024 to assess whether recommendations had been used and how participants perceived both recommendations and process.

According to Björnqvist et al. (PLOS ONE, 2026), the authors used M365 Copilot (GPT-5) to translate Swedish map statements to English and to recreate visual structure, noting explicitly that the LLM was not used for data generation, analysis, or interpretation.

Direct participant feedback in PLOS ONE (Björnqvist et al., 2026) illustrates outcomes: "I think this became a bit more concrete, " said Interviewee 2, and "If they [the recommendations] had been in place when the pandemic came, we would have had it easier during the pandemic, " said Interviewee 7.

Findings Snapshot

Date	Metric	Value	Implication
3 Apr 2023 – 12 Jun 2023	Workshops	4 workshops with 15 participants	Enabled iterative mapping, validation, recommendation drafting (PLOS ONE, 2026)
April–June 2023	Recommendations produced	10 participant-developed recommendations	Targeted central risks identified in maps (PLOS ONE, 2026)
Spring 2024	Follow-up interviews	10 of 15 participants interviewed	Used to assess perceived use and impact of recommendations (PLOS ONE, 2026)
20 Aug 2026	Publication	PLOS ONE case report published	Provides peer-reviewed evidence for the method and practical lessons (PLOS ONE, 2026)

Implications for Qualitative Researchers and Evaluators

Implications: The PLOS ONE case suggests participatory mapping plus iterative validation can increase process use and generate implementable recommendations, but clarity and facilitator involvement matter for actionability.

Define "use" before follow-up: Björnqvist et al. (PLOS ONE, 2026) highlight that interviewee responses varied because the study lacked a single pre-defined operational definition of recommendation use.
Balance involvement and quality: Björnqvist et al. (PLOS ONE, 2026) recommend facilitator edits or a brief follow-up validation session to increase recommendation clarity while preserving participant ownership.
Shorten where needed: Björnqvist et al. (PLOS ONE, 2026) note four workshops produced fatigue for some participants and suggest testing a two-workshop variant for efficiency without losing core causal-mapping benefits.
Use AI for translation and transcription verification: Björnqvist et al. (PLOS ONE, 2026) used an LLM for translation and map recreation but verified outputs manually; the study recommends AI as a tool for efficiency, not interpretation.

How Evidano Helps: AI-enabled Qualitative Research for Participatory Evaluations

Problem: Slow synthesis of workshop maps and transcripts

Solution: Evidano streamlines transcription, translation, and thematic synthesis so evaluators can convert workshop causal maps and interview transcripts into coded findings faster.

Contextual link: See Evidano features for thematic coding, cross-segment analysis, and visualizations that mirror the causal mapping outputs described in PLOS ONE (Björnqvist et al., 2026).

Problem: Unclear or vague participant-written recommendations

Solution: Evidano supports collaborative coding and iterative refinement workflows so facilitators can produce clearer, actionable recommendation drafts and then re-share for participant validation.

Contextual link: Use Evidano speech-to-text to get accurate transcripts from recorded workshops and interviews, matching the verification step Björnqvist et al. applied to AI outputs (PLOS ONE, 2026).

Problem: Tracking which recommendations were used over time

Solution: Evidano offers cross-segment and frequency analysis that helps evaluators tag recommendations, log implementation status, and generate concise follow-up reports for managers and stakeholders.

Practical note: Björnqvist et al. (PLOS ONE, 2026) report that participants perceived varying levels of recommendation use; Evidano makes those differences auditable with traceability to quotes and timestamps.

FAQ: AI qualitative analysis participatory evaluation

Can strategy mapping software be evaluated using AI-enabled qualitative methods?

Yes. The PLOS ONE case demonstrates that strategy mapping outputs can be translated, transcribed, coded, and summarized with AI support and human verification (Björnqvist et al., 2026).

Supporting detail: Björnqvist et al. (PLOS ONE, 2026) used an LLM to translate Swedish maps to English but manually verified all outputs and limited AI use to non-interpretive tasks.

How much data and how many sessions did the PLOS ONE case use?

Answer: The PLOS ONE case involved 15 participants across four workshops held between 3 April 2023 and 12 June 2023, producing ten recommendations (Björnqvist et al., 2026).

Supporting detail: The authors followed up with ten semi-structured interviews in spring 2024 to assess use and perceptions of the recommendations (Björnqvist et al., 2026).

What are common pitfalls to avoid when combining participatory mapping with AI?

Answer: Avoid treating AI outputs as final interpretations; Björnqvist et al. (PLOS ONE, 2026) caution that AI should be used for translation and layout but not for analysis without human verification.

Supporting detail: The PLOS ONE authors manually reviewed LLM translations and explicitly stated the LLM was not used for data generation, analysis, or interpretation (Björnqvist et al., 2026).

How can I measure whether recommendations from a participatory evaluation were used?

Answer: Predefine operational criteria for "use" and collect both documentary evidence and participant reports; the PLOS ONE case found that lack of a common definition produced divergent interview responses (Björnqvist et al., 2026).

Supporting detail: Björnqvist et al. (PLOS ONE, 2026) recommend clear definitions and structured follow-ups to reduce variance in perceived use across stakeholders.

Conclusion & Next Steps

The PLOS ONE case (Björnqvist et al., 2026) shows that participatory strategy mapping delivered ten recommendations and stakeholder learning when run across four workshops in April–June 2023 and followed up in spring 2024.

For evaluators, the practical lesson is to pair participatory mapping with precise definitions of use, facilitator-supported recommendation refinement, and audited follow-ups so recommendations become actionable, as discussed in PLOS ONE (2026).

If you run participatory evaluations and want to accelerate transcription, translation, thematic coding, and traceable follow-up reporting, Evidano can replicate the PLOS ONE workflow at scale and with audit trails.

Next step: Try Evidano for free to import transcripts, map and code recommendations, and produce follow-up reports that preserve participant quotes and implementation status.

Topics

AI qualitative analysis participatory evaluation
participatory evaluation strategy mapping
AI-enabled qualitative research
strategy mapping software evaluation

Keep reading

Browse all articles