Piloting Cogniti: Exploring educator-designed AI agents for learning

Why Cogniti?

As generative AI becomes increasingly embedded in higher education, universities are exploring how to use it in ways that genuinely enhance learning rather than simply automate tasks or provide answers.

To support this exploration, Teaching and Learning Innovation is piloting Cogniti, an AI platform developed by the University of Sydney and now used across several Australian universities. Cogniti enables educators to create and deploy AI agents without coding by using plain-language instructions, subject resources and carefully designed guardrails. These agents can be embedded directly into Canvas, providing contextual support where students need it most.

The pilot aims to explore how educator-designed AI agents can support student learning at scale while remaining aligned with subject objectives, assessment requirements and teaching practice. It also provides an opportunity to build institutional capability and identify effective approaches to AI-enhanced learning that can be shared across the university.

Expressions of interest to participate in Cogniti Semester 2 2026 pilot have now closed. We will share the findings and outcome of the pilot ahead of Semester 1, 2027.

Why use an AI agent in your teaching?

An AI agent is a purpose-built conversational assistant designed to support a specific learning goal or educational task.

Unlike general-purpose AI tools, an AI agent is configured by educators using subject-specific instructions, resources and behavioural guidelines. This allows the agent to operate within clearly defined boundaries and respond in ways that align with the learning context.

For example, an educator might design an AI agent to provide:

• Socratic guidance - encourages students to think critically by asking questions, prompting reflection and guiding reasoning rather than providing answers.
• Scalable formative feedback - enables students to test ideas, refine arguments and check their understanding, with timely feedback available whenever they need it.
• Scaffolded practice - supports learning through worked examples, hints and checks for understanding, with support gradually reduced as confidence and competence develop.
• Authentic role-play - simulates interactions with clients, patients, historical figures or interview panels, allowing students to apply knowledge and practise professional skills in a low-stakes environment.
• Accessible, on-demand support - provides a patient, non-judgemental learning companion that students can engage with at their own pace, helping to improve access and participation.

Embedding these experiences directly within Canvas allows students to engage with AI as part of their normal learning environment, reducing barriers to access and creating a more integrated learning experience.

Principles for designing AI agents for learning

The pilot is guided by the principle that effective AI integration begins with pedagogy.

AI agents should be intentionally designed to support student learning by encouraging reasoning, reflection and engagement, while maintaining the central role of educators and human judgement.

Key design principles include:

Learning first

AI agents should be designed with a clear educational purpose and aligned with the subject's intended learning outcomes, learning activities and assessment. Their purpose, role and appropriate use should be clearly communicated to students. Agent interactions should promote inquiry, critical thinking, problem-solving and reflection to support meaningful learning.

Scaffold, don't solve

Rather than providing completed answers, agents should guide students towards understanding through questioning, feedback, hints and progressively reduced support. The goal is to build learner confidence and independence over time.

Academic Integrity

Educators should clearly communicate when and how AI can be used, ensuring alignment with university policies and academic integrity expectations.

Establish clear guardrails

The behaviour of AI agents should be guided by carefully designed instructions, approved resources and defined boundaries. This helps ensure responses remain accurate, relevant and aligned with the subject context.

Evaluate and improve

The pilot provides an opportunity to evaluate how AI agents influence student engagement, learning experiences and feedback. Insights gained will help identify effective practices, inform future development and support responsible adoption of AI across the university.

For more information on Cogniti, please see the Cogniti information and help pages.

To talk through the effective use of Cogniti with our learning designers, please book a consult.