Evaluation at a Crossroads: Personal Reflections from EES2026
Updated: 4d

Attending the 16th European Evaluation Society (EES) Conference in Lille, France, was a first for me. I arrived with three goals: to explore new methodological and thematic developments, share lessons from our work at Key Impact, and better understand where the MEAL profession is heading.
The conference brought together a diverse group of professionals: evaluation commissioners, academics, practitioners, representatives of consulting firms, independent consultants, and in-house MEAL specialists. Despite our different backgrounds, I was struck by a shared sense of purpose and urgency around some questions:
What is the role of evaluation when AI can increasingly perform tasks traditionally undertaken by evaluators?
How can evaluation remain relevant as democratic spaces and resources shrink?
How can we make our work more inclusive and, ultimately, more influential in decision-making?
These questions formed a common thread across many of the sessions I attended. What I found encouraging was the positive attitudes with which participants approached them, combining honest reflections on the challenges with a willingness to rethink established practices.
Four sessions particularly stood out for me.
1. Living the Tensions: Responsible AI Evaluation Case
Unsurprisingly, AI featured prominently throughout the conference. While some discussions felt repetitive, I particularly appreciated this session. One reassuring message from Julian King was that we are still far from a one-minute evaluation. More importantly, the discussion reinforced the idea of establishing a complementary relationship between AI and human evaluators, in which each contributes what it does best.
AI can automate routine tasks, rapidly process large volumes of information, support data analysis, and assist with coding and pattern recognition. Human evaluators, however, remain essential for exercising evaluative judgment, building relationships and trust, navigating ethical dilemmas, and bringing creativity and contextual understanding to the process.
My main takeaway was that we should allow AI to challenge us as evaluators, not simply make us faster. Beyond improving efficiency, AI can help us question our assumptions, identify gaps in our reasoning, explore alternative interpretations, and recognize weaknesses in our analysis. Another thought that resonated with me came from Mhairi Aitken during the second plenary discussion. She expressed her wish that we move on to discussing the more “boring stuff” about AI: the specific tools evaluators can use, how to apply them responsibly, and when they genuinely add value to the process.
I share this sentiment. While debates about AI's opportunities and risks remain essential, I would welcome more practical exchanges about its application in real evaluation assignments. Perhaps it is time to move from debating whether AI will transform evaluation to exploring how it can systematically improve our work without compromising the human dimensions.
2. From Stories to Data: Capturing Lived Experience, Sentiment and Language Nuances
This session left a strong impression on me. Discussions about inclusive evaluation often focus on bringing local voices into the process. This session went a step further by highlighting the importance of understanding not only what people say, but also their sentiments, emotions, identities, attachments, and relationships that shape their experiences. These dimensions influence motivation, trust, engagement, and ultimately the effectiveness and sustainability of interventions.
The session reinforced my belief that inclusion is not simply about consulting more and different people, but about creating opportunities for diverse actors to shape how evaluations are designed, conducted and interpreted. This includes building the capacity of local evaluators, recognizing their contextual knowledge, and giving them meaningful opportunities to contribute their judgements throughout the evaluation process. But there is also another dimension: the evaluator's own position. As evaluators, we are often the narrators of the programs we assess. We select which evidence to emphasize, how to connect different perspectives, and how to communicate the story of change. These choices are inevitably shaped by our own assumptions, professional and cultural backgrounds.
Being aware of that matters, and continuously reflecting on it is essential. Whose perspectives are we prioritizing? What assumptions guide our interpretations? What might we be overlooking? For me, inclusive evaluation requires both openness to others' perspectives and a willingness to critically examine our own.
3. Seeing Evaluation Differently: Visualization as Method in Theory, Program Design and Evidence Synthesis
This was one of the most thought-provoking sessions I attended during the conference. I was particularly struck by Bianca Montrosse-Moorhead's argument that visualization goes far beyond presenting results. How we visually represent processes, evidence, and findings influences who can understand them and ultimately act on them.
This message resonated strongly with our experience at Key Impact. Over the past few years, we have invested considerable effort in improving how we communicate visually with our clients throughout the evaluation process. We increasingly recognize that visual communication is not merely about producing appealing reports or presentations. It can simplify complex designs, clarify methodological choices, make analytical frameworks accessible, facilitate collective interpretation, and support decision-making. Yet the session made me realize how much further we could go. What if we treated visualization not as the final step in communicating findings, but as an integral part of evaluation the entire evaluation process?
I would have welcomed more opportunities at EES2026 for practitioners to exchange concrete examples, tools, and challenges related to visual communication throughout the evaluation cycle. After all, even the most rigorous evaluation risks having limited influence if its process and findings are difficult for intended audiences to understand and use.
4. Future-Informed Evaluation: Navigating Complexity and Uncertainty in a Rapidly Changing World.
In the final session that stayed with me, the discussion explored the potential of integrating foresight approaches into evaluation, particularly in complex and rapidly changing contexts.
How can we ensure that our findings and recommendations remain useful when the context itself is changing? Foresight approaches offer a promising way to answer this question. By exploring alternative future scenarios and examining how contextual developments might affect outcomes, evaluators can help organizations anticipate challenges and make more adaptive decisions.
This is not about predicting the future. Rather, it is about acknowledging uncertainty and considering how different choices might perform under different scenarios. For me, integrating foresight is a compelling way to strengthen the relevance of evaluation. Instead of simply recommending what should change based on past experience, we can help organizations consider which strategies may remain effective in different futures, which assumptions need monitoring, and when adaptation might be necessary.
Looking Ahead
EES2026 reminded me that evaluation is not simply about methods and rigour. As a field, it must continuously question its assumptions, adapt its practices, and demonstrate its value. The sessions and discussions throughout the conference strengthened my belief that the future of evaluation may depend not only on producing rigorous evidence, but increasingly on how we generate, interpret, communicate, and use that evidence.
More importantly, I came back from Lille with some concrete ideas for our work at Key Impact: to experiment more deliberately with AI as a critical tool, deepen the inclusion of various perspectives and reflexivity in our processes, make visual communication a more purposeful part of our work, and explore how foresight can inform our analysis and strengthen our recommendations.


Comments