Plato AI Pilot at the University of Liverpool

Posted on: 16 January 2026 by Will Moindrot and Laura Blundell in General

We have just completed a pilot of Plato where a sample of 12 first-semester modules trialled the solution within their Canvas environment. Across the University and the sector more broadly, there has been a palpable nervousness regarding the embedding of AI tools into long-standing systems like Virtual Learning Environments. Plato allowed us to explore this space in a positive, controlled, and highly visible way.

Plato AI – Pilot Background (3:55mins)

Strategic Affordances vs. Generic Alternatives

Our previous experimentation with "disparate" AI tools and generic Custom-GPTs often faced hurdles with student friction and pedagogical trust. Plato provided a significant step forward due to several key affordances:

• Seamless VLE Integration: Unlike external chatbots that require separate logins and context-switching, Plato’s presence directly within the Canvas VLE removed the "pain" of system access. This frictionless experience was a major driver of our high engagement levels.
• Granular Customisation as Pedagogical Extension: The ability to customise the tool at a programme or module level is a "huge" benefit for staff. It shifts the AI from a generic tool to a specific extension of a lecturer's pedagogy. Staff can shape the AI's "persona" and define its role in student interactions with a level of granularity we found missing in other systems.
• Student Confidence through Faculty Oversight: A significant draw for students was the knowledge that the tool was developed by their actual lecturer—the same person designing their curriculum and setting their assessments. This created a level of trust that generic personal AI setups cannot match.
• Staff Buy-in and Legal Security: The clarity provided by our legal contract regarding content ownership and handling was vital. This gave staff the confidence to get on board and, crucially, allowed them to reassure teaching colleagues about the safety of their intellectual property.

Impact, Analytics, and Data Sovereignty

The platform provided a "ready-to-roll" solution that demonstrated to students that the University is proactively and safely engaging with AI.

• High Engagement: Over 1,335 students signed up (representing 53% of the trial cohort), asking more than 22,576 questions. This provided the equivalent of 1,880 hours of support, with 25% of queries occurring outside of business hours.
• Actionable Analytics: Beyond simple usage metrics, the aggregate data allowed us to drive immediate teaching enhancements. Students, who are often weary of "feedback fatigue," derived great satisfaction from seeing their interactions with Plato lead to visible course improvements.
• Data Sovereignty: From an institutional perspective, it is strategically vital that we capture and use this interaction data to improve our own teaching, rather than allowing that value to be harvested by external companies with whom we have no formal relationship.

(Further evaluation data will be made available when our project group, led by Dr Maria Limniou, reports later this year).

From a Learning Technologist’s perspective, the Plato team was incredibly agile and responsive, tweaking settings within hours to suit our needs. Overall, we’ve had a really positive experience; the platform is a powerful resource for scaling curriculum-aligned support while maintaining pedagogical integrity and institutional control.

Ultimately, the pilot of Plato has shown us the way for how AI can be effectively embedded within the learning space to derive new value from high-quality pedagogy.