Teachable Agents is a learning technology that draws on the social metaphor of teaching a computer agent to help students learn (AAA Lab, Stanford).

When students use Knowalla, they engage with a teachable agent named Olli. Olli is trained to “learn” from the student and to use only the knowledge the student has provided.

Knowalla is built on learning by teaching, a research-backed method that helps students:

• Strengthen knowledge (Duran, 2017; Fiorella & Mayer, 2014)

• Improve recall & transfer of understanding (Nestojko et al., 2014)

• Organize thoughts & spot gaps in comprehension (Chase et al., 2009)
  
  

The Protégé Effect: Students feel more accountable when teaching an AI “teachable agent,” boosting motivation, persistence, and effort (Chase et al., 2009).

Knowalla uses concept mapping to help students organize knowledge visually and connect ideas meaningfully (Novak, 1990; Daley & Torre, 2010).

When the watches their teachable agent apply what it has learned to answer questions, they receive recursive feedback.

Definition: Recursive feedback occurs when a teacher observes their students apply what they have been taught (Okita & Schwartz, 2013).

In the case of Knowalla, the character’s performance provides the user with indirect feedback about their own understanding of the material

Recursive Feedback In the Knowalla learning pattern:

1. The AI agent answers questions using the student’s concept map.

2. Students identify errors, revealing gaps in their own understanding.

3. Mistakes are low-risk—students can adjust their thinking without damaging confidence (Kluger & DeNisi, 1996; Chase et al., 2009).

By refining concept maps and re-testing the AI, students actively manage their own learning (Butler & Winne, 1995; Zimmerman, 2000).

How Knowalla Supports Self Regulated Learning (SRL)

1. Students monitor their understanding by checking Olli’s performance.

2. Students adjust strategies to fill knowledge gaps.

3. Students practice metacognition—planning, tracking, and refining learning (Schunk, 2012; Panadero, 2017).

This cycle builds independent, confident learners.

Why it works:

Teachers benefit from the platform’s formative assessment capabilities (Black & Wiliam, 2009; Hattie & Timperley, 2007). As students build and adjust concept maps, educators can observe common misconceptions and adapt instruction in real time. In this way, Knowalla leverages a feedback loop that is quick and actionable, improving both student learning outcomes and instructional efficiency (Winstone et al., 2021).