Causal Map Workflows:
Use AI to make sense of mountains of texts — rigorously and verifiably
What is Workflows?
Workflows is a tool used by Causal Map Ltd to make sense of large amounts of text data (interviews, reports) supported by generative AI.
At the moment, use of Workflows is not open for public use. Instead, we use it internally to provide verifiable and transparent analysis and reporting to clients. Clients can then access their workflows in the app to view and, if they wish, verify each step.
Which questions can Workflows help answer?
- All of the Questions You Can Answer with Causal Mapping, for example:
- Causal pathways
- New, emerging and unexpected factors and outcomes
- Project impact through the eyes of stakeholders and other sources
- Contribution of projects to outcomes
- QuIP: Synthesis of Qualitative Impact Protocol evaluations
- OH: Synthesis of Outcome Harvesting processes
- MSC: Synthesis of stories collected for Most Significant Change processes
- Comparisons between groups and time-points
- Vignettes: which stories are both remarkable and frequent?
- More generally, any kind of question you can answer by analysing text systematically, reproducibly and verifiably, page by page or paragraph by paragraph:
- Stakeholders’ mental models
- Social Network Analysis
- Thematic Analysis
What are the benefits?
- Scalability: Process large amounts of text data quickly and efficiently.
- Transparent & Verifiable Results: Based on a clear, reproducible audit trail.
- Include more voices: Not just small samples of the easily accessible ones.
For Evaluators and Contractors
- Enhanced Bids: Offer state-of-the-art techniques to make your proposals more competitive.
- Time & Cost Savings: Reduce manual coding and analysis efforts.
- Reproducible Workflows: Guarantee that analysis steps are documented and repeatable.
Why is Workflows better than tools like ChatGPT or NotebookLM?
- break down big, vague tasks into small, easy ones so that we don’t leave the AI to make big judgements all on its own
- use a clear set of basic tools for each of those small, easy steps
- speed up and automate the application of hundreds of steps like prompts and filters
- manage and keep track of your work so you know exactly what you did last year, and can reproduce the same results in the same way.
Why Causal Map Ltd?
- Innovation in qualitative analysis and reporting
- Broad experience in evaluation and social science research.
Consultancy Services
In addition to our user-facing tools, we offer bespoke consulting services that leverage AI to streamline causal mapping projects from start to finish. Our expert team will work closely with you to:
- Develop research questions that get to the heart of your investigation
- Process and code your qualitative data using Causal Map Workflows.
- Deliver insightful, visually compelling reports that directly address your research questions
With the power of AI and our deep causal mapping expertise, we can uncover the causal stories in your data with unparalleled efficiency and depth of understanding. To explore how our consulting services can advance your research goals, get in touch with our team today.
FAQs
What does it cost?
We charge by day of our time, £695+VAT/day.
Most of the work isn’t using the app. Most of our time is spent finding out what you want to know and what findings can help you, and iterating our reports until you are happy.
What are the caveats?
Like humans, the AI processor makes mistakes and needs to be monitored. If you want an overall picture, the "noise" it can produce probably won't matter and you can get very good results almost out of the box. However at the moment, if you want to be sure that each individual piece of coding is correct, you'll probably need to tweak some proportion of the coding it produces.
Why causal mapping?
At Causal Map Ltd., we promote causal mapping as a really useful and effective way of processing text which can be applied with surprisingly little adjustment to a great variety of evaluation tasks. Our Workflows software provides ready-made causal mapping steps, which can be used alone or seamlessly alongside other text processing tasks like thematic analysis. And the software can of course produce graphical causal maps. As the software can also be used without any causal mapping steps or visualisations, it could In principle be seen as also occupying a similar space to any of the market-leader CAQDAS applications which now offer AI assistance, but with more explicit, transparent workflows. However it is not our intention to compete directly with this kind of general-use software.
What is wrong with using AI as a "black box"?
At Causal Map Ltd, we use AI to both collect and analyse qualitative data. 🤖 We believe that the use of AI should also follow published guidelines to ensure transparent and valid results.
We are not comfortable with procedures which rely on giving the AI the freedom to make evaluative judgements. So we do not, for example, ask the AI “what are the main or most important causal stories in the document”. We do not ask the AI to make summaries. Instead, we use the AI only as a tireless low-level assistant to exhaustively and transparently identify large numbers of individual causal claims (ideally, every single one of them) within the texts.
🔍 We follow established qualitative social science procedures for coding, with the aim that each step of the process from coding to analysis is transparent and verifiable. We don’t rely on AI to simplify the coded data or decide, for example, which themes are most important.