Model evaluation workshop for leaders

​Continuous evaluation and iteration is the only way to develop a successful AI product. To do it well, we need to be clear about what evaluation means.

​In the social sector, evaluation isn't just about model accuracy. We must assess whether the AI product or service drives positive, measurable change in people’s lives.

​The 4-level evaluation framework by Agency Fund serves as a guide to make that kind of evaluation not only possible, but practical.

​Level 1: Does the AI system perform as intended?

​Level 2: Does the product engage and retain users?

​Level 3: Does the product impact users’ thoughts, feelings, knowledge, and behavior in a way that moves towards the development outcome?

​Level 4: Do users with access to the product improve development outcomes?

​Each of these levels require separate deep dives to be understood properly.

​In this workshop, we will cover: Level 1: Does the AI system perform as intended?

​Evaluating AI systems is typically made to sound more complicated than it needs to be with leaders lacking the vocabulary to communicate with their team often feeling confused about where they can add value. This workshop aims to clear the cutter and bring you back to the driving seat so that you can help your team with the clarity they need to move with high velocity, and not just speed.

​It is intended for leaders working on socially impactful challenges who don't stop at measuring engagement and care about driving meaningful ships in the final developmental outcome - be it providing quality care, improving learning outcomes or anything else.