Four strategies for implementing custom AIs that help students learn, not outsource

Adobe Stock, used under licence, generated with AI

The launch of Cogniti has opened new opportunities for teachers at the University of Sydney to design customised AI agents that align with unit outcomes, assessment tasks, and discipline-specific ways of thinking. When designed and implemented well, Cogniti agents can guide students through practice tasks, model disciplinary reasoning, and provide personalised feedback that enhances, not replaces, their learning. How we introduce and contextualise these agents is one of the most important factors in determining whether students benefit from them.

Left unexplained, students may treat Cogniti like another version of ChatGPT, a tool to “get the answer quickly” or outsource their work. If we instead intentionally frame our Cogniti agents in our teaching, model how to use them, and highlight their unique value, students are more likely to use them productively, to consolidate understanding, practise skills, and deepen their thinking.

For Cogniti to be taken seriously, it needs to be woven into the structure of your unit and its delivery, both in class and on Canvas, rather than left on the side. This article shares practical strategies for implementing Cogniti in your teaching so that students:

  • understand the context and purpose of the agent,
  • know how to interact with it effectively, 
  • perceive its value as a learning tool over any other available AI chatbots, and
  • engage in reflection and feedback.

In this post, we discuss how to introduce and integrate Cogniti agents into the learning environment so students understand their context, interact effectively, and see their value as customised learning companions.

In this post, we share four strategies to help introduce and integrate Cogniti in your teaching so that students understand their context, interact effectively, and see their value as customised learning companions.

Students need to see Cogniti as a trusted learning partner, not just another chatbot. The way we introduce agents and model their use makes all the difference.

Sydney Unit Coordinator, 2025

1) Give students the context

Students need to know what Cogniti is, and how it differs from generic AI tools. While ChatGPT is a general-purpose assistant, Cogniti agents are curriculum-aligned, customised, and framed for learning. They have been built by their own teacher to reflect unit-specific content, learning outcomes, and academic expectations. When introducing your Cogniti agent in class and/or on Canvas, help students see the agent as part of their learning environment, not just another optional add-on.

Consider:

  • Explaining its role: Define the purpose of the agent–is it meant to help students practice applying something or preparing for an assessment?
  • Linking to outcomes: show how the agent supports key learning outcomes or assessment preparation.
  • Clarifying boundaries: emphasise that the agent supplements teaching and feedback, and does not replace attending class, engaging with readings, or completing tasks independently.
  • Deploy in Canvas where it matters: The most effective placement is usually within a Canvas page or assignment, linked directly to the relevant module. For example, if students are preparing a group presentation, positioning the agent in that assignment space signals its direct relevance.

2) Model how to interact with the agent

Students may not naturally know how to use Cogniti effectively. A brief in-class demonstration can make a big difference. Consider:

  • Live demo: Ask the agent a question relevant to the week’s topic and show how it responds. Highlight useful features, like asking follow-up questions or requesting examples.
  • Model its use during live classes: One of the best ways to demonstrate how students can best engage with Cogniti is to perform a live demonstration during classes that relate to its purpose.  
  • Pre-tutorial practice: Encourage students to try the agent before a tutorial, then bring their questions or outputs into class discussion.
  • Good vs. poor prompts: Compare a vague prompt (“Explain this topic”) with a specific one (“Can you give me a step-by-step explanation of how X relates to Y in this week’s case study?”).
  • Active learning: Encourage students to use the agent for practice rather than answers, e.g., generating quiz questions, testing draft ideas, or checking their reasoning.

By making the invisible visible, you give students a toolkit for self-regulated interaction with AI, a skill that will serve them well beyond your subject.

3) Emphasise the value for learning

The most powerful way to prevent misuse of Cogniti (e.g., outsourcing thinking to AI) is to help students perceive its unique value for their learning. This involves:

  • Positioning Cogniti as a practice space: “The agent is here to give you low-stakes practice so you can test your understanding before assessments.”
  • Highlighting personalisation: Unlike generic AI, Cogniti is designed with your unit, tasks, and expectations in mind.
  • Linking to authentic skills: Frame Cogniti as a way to practise disciplinary thinking (e.g., clinical reasoning, essay structuring, design critique), not just content recall.
  • Reinforcing academic integrity: Be explicit that Cogniti supports students to develop their ideas and skills, not replace them.
  • Extend opportunities for practice: Cogniti can provide low-stakes rehearsal in large classes where personalised feedback is limited.

When students understand this value, they are more likely to engage meaningfully rather than seeing the tool as a shortcut (Selwyn, 2022).

Doctor of Veterinary Medicine student’s experience of a mentoring agent called Centaur the Mentor:

“It [the mentoring agent] provided me with great suggestions, and I understand what you meant by the Centaur leading me to think through questions.”

VETS6210 student, 2024

4) Invite reflection and feedback

Learning is best supported when students are prompted to reflect on their experience. For example, you might ask them to discuss or write a short reflection on:

  • What they learned by engaging with the agent.
  • How the agent’s responses helped (or didn’t help) them understand a concept.
  • How they might use the agent differently next time.

Gathering student feedback also helps you refine the agent and its integration in your unit. For example, students may highlight unclear instructions or suggest ways the agent could better support their preparation. Adding a simple feedback opportunity to the page where students interact with the agent, using tools such as SRES or Qualtrics, can help you understand your students’ experience (see this page for a how-to guide).

Student feedback on their use of an agent:

I was actually one of the people that were unenthusiastic about the use of AI when others were obsessed over ChatGPT. But now I understand what being able to utilise AI means and the importance of digital intelligence.

VETS6201 student, 2025

Final thoughts

Cogniti offers an opportunity to intentionally embed generative AI into learning and teaching at the university. However, the technology itself is not the solution–how we implement it is what matters most. By giving students clear context, modelling effective interactions, highlighting the value of Cogniti for learning, and embedding it into your unit design, you can help them actively engage with AI in ways that support critical thinking, academic integrity, and lifelong learning skills.

Cogniti agents are more than just chatbots. Implemented thoughtfully, they can become trusted learning companions, helping students practise, reflect, and grow.

Find out more

As you start exploring the use of Cogniti in your teaching practice, you may find the following helpful: 

More from Kria Coleman

Group work – rich, meaningful and transformative learning

Group learning is an increasingly popular approach to assessment; it provides opportunities for collaborative...
Read More