🌻 AI-assisted causal mapping. Uncovering causal pathways with INTRAC
PDF

11/12/2025

We partnered with INTRAC on a project for the SCC (Strengthening Civil Society) programme. This programme is led by a consortium of four organisations: PAX, ABAAD, DefendDefenders and Amnesty International Netherlands, with INTRAC serving as the MEL partner. INTRAC is a non-profit organisation with a deep commitment to understanding and improving the practice of civil society worldwide.

The SCC programme is a complex, multi-year initiative spanning 13 countries. The INTRAC team was using Outcome Harvesting but found themselves struggling to get the 'big picture'. They needed a systematic, yet flexible, way to connect the dots in a large body of qualitative data to answer critical questions about the programme's overarching causal pathways of change.

The core challenge was translating a high volume of text into high-quality, actionable analysis. Alastair Spray, Senior Consultant at INTRAC, led the causal mapping component of the End Term Evaluation. He worked alongside a core team of four, supported by two other colleagues who provided input into the wider evaluation. As Alastair noted:

*"The client had been pursuing usual evaluation approaches... and had found Outcome Harvesting useful, but they felt that they couldn't understand the big picture, for the whole programme but also for specific countries too."

The team needed to rigorously analyse the data to answer questions such as:

We used Causal Map's consulting service and our AI-assisted approach to bring clarity and rigour to this complex qualitative data, creating a process that was both systematic and reproducible for INTRAC.

Given the sensitive nature of the SCC programme, we worked closely with the team to ensure all data was thoroughly anonymised before the mapping and analysis began.

Our approach#

  1. Data input: We processed comprehensive data sources, including Harvested outcomes data from the 13 countries across four programme years, and Narrative Theory of Change Review documents (typically 8–15 pages long) from partner organisations.
  2. AI-Assisted coding: The Causal Map team was responsible the core coding. We used a radical zero-shot approach with our AI (Gemini 2.5 Pro), which was instructed to invent its own codes based purely on the text. This process quickly identified 5,430 causal claims from the respondents, which were also auto-coded for sentiment (positive/negative links).
  3. Collaborative Analysis: We provided a preliminary analysis report and materials, along with specific training for Alastair on the cm4 platform. This allowed the INTRAC team to engage directly with the maps and dig deeper into the coding when writing their final report. This hands-on support was crucial for building confidence, which Alastair highlighted:

    “The team (Steve and Gabriele) are incredibly helpful, and good at training you up on how you might want to use it to interrogate your data. My advice would be to learn by doing, try to get started and then ask for help as you need it, rather than worrying about getting it right from the start.”

The use of Causal Map for this analysis generated a powerful set of visual and analytical outputs, allowing INTRAC to move from raw data to a better understanding of 'what causes what' in their programme.

INTRAC MAP

The project provided INTRAC with a robust, visual evidence base that articulated how stakeholders perceived change in the SCC programme, helping them to move from raw data to a deeper understanding of 'what causes what' in their complex field.

See what Alastair has to say: