Is AI Making SROI Easier? Yes—But It Won't Human Judgement

AI support and human judgement in SROI, showing how AI can summarise data, draft insights and organise evidence while people assess materiality, validate findings and make decisions

With the rapid advancement of Large Language Models (LLMs) such as ChatGPT, Claude and Gemini, one question which came up was: "Can AI support an SROI?"

SROI is an evidence-based, stakeholder-centred methodology. While AI is becoming remarkably good at processing information, it cannot replace the judgement, ethics and stakeholder engagement that lie at the heart of a credible SROI.

Where AI is already proving valuable is in reducing the time spent on technical and administrative tasks. For example, AI can help practitioners:

• conduct rapid literature reviews;

• identify potential financial proxies from published research;

• summarise government reports and datasets;

• draft and refine Theory of Change narratives;

• code qualitative interviews and identify emerging themes;

• synthesise stakeholder feedback across hundreds of interviews;

• suggest alternative financial valuation methods;

• prepare tables, charts and reports; and

• improve the clarity and consistency of final deliverables.

Tasks that once took days can often be completed in hours, allowing practitioners to spend more time interpreting findings and engaging with stakeholders.

However, there are several aspects of SROI that AI cannot replace.

It cannot decide which outcomes are genuinely material to a community. It cannot facilitate a participatory Value Game or understand local cultural nuances. It cannot determine whether a financial proxy is appropriate for a particular context simply because it found one in a database.

Nor can it make methodological judgements about attribution, deadweight or unintended consequences without human oversight.

Perhaps most importantly, AI cannot build trust.

The quality of an Social return on investment (SROI) depends on meaningful conversations with stakeholders, careful listening, and the ability to understand why an outcome matters, not just whether it occurred. Rather than replacing practitioners, AI is likely to change the nature of their work.

Less time will be spent searching for evidence, formatting reports and coding transcripts. More time can be devoted to asking better questions, interpreting complex findings and supporting organisations to make better decisions. 

It is giving good practitioners the opportunity to become even better. The future of SROI is therefore unlikely to be AI versus people. It will be AI-assisted practitioners who combine technical expertise with empathy, critical thinking and stakeholder engagement.

Watch the full webinar here: https://swiy.co/WebinarSROI.

https://www.youtube.com/watch?v=b7glYaklW6w

If you're looking to build your first SROI, the following open-access resources are the best place to start.

Resource	What it includes
Guide to SROI – Social Value International	The definitive step-by-step guide to conducting an SROI, from stakeholder engagement and Theory of Change to valuation, discounting and reporting.
Value Map Template – Social Value International	The standard template is used to map stakeholders, outcomes, indicators, financial proxies and impact adjustments across any SROI study.
Standards and Guidance – Social Value International	Detailed guidance on applying the Principles of Social Value, including materiality, stakeholder engagement, valuation and assurance.
Social Return on Investment (SROI) Sample Reports	Case studies, technical papers and practical examples across different sectors.
Global Value Exchange	A searchable database of indicators, outcomes and financial proxies that can support valuation, while always requiring contextual adaptation.