AI — Pallara

Purpose-built AI

Most platforms added a chatbot. Pallara built an AI layer.

Tutoring for every student. Plain-English answers for every staff member. Early warning before problems become withdrawals. All governed from your own control centre, with learner-scoped access and the models your provider chooses.

Book a walkthrough Meet Aria

For learners · 90 seconds

Know what comes next.

See how learning, Aria, notebooks, assessment, wellbeing and mobile access form one clear learner journey.

Architecture

A layer of the platform, not a feature on the side.

Every agent in Pallara — Aria, the staff assistant, at-risk scoring, content design — runs through one AI engine built into the platform itself. And that engine is provider-agnostic: any agent can connect to any model, from any provider.

When a better model ships, you switch by configuration — same agents, same policies, same logs. No migration, no waiting for us, no vendor lock-in. Your AI stays current as the field moves.

Your agents
Aria — tutor Staff assistant At-risk scoring Content designer
Pallara AI engine policies · approvals · logging · privacy
Any model — swapped by configuration
OpenAI Anthropic Google Open-source Locally hosted

For students

Aria, the personal tutor every student gets.

Aria is grounded in the student's own course material, with inline citations back to the lesson. She explains concepts, generates study packs, revision sheets and quizzes — including from the student's own notebooks.

Guides. Never writes the answer.

On drafts, Aria coaches Socratically against the rubric — asking the questions that improve the work, not producing it. Academic integrity is a design constraint, not a policy note.

A Aria NZ Cert in Business L5 · draft coaching
Here's my draft answer for Task 2. Is it good enough to submit?
You've described what the marketing mix is — but the rubric asks you to apply it. Rubric · Criterion 2

Look at your paragraph on price: which pricing strategy is your case-study café actually using, and why might they have chosen it? Try answering that in two sentences, then we'll look at promotion.
Show me the rubric Quiz me on pricing Make a revision sheet

For staff

Ask your data. Get an answer, not a report request.

Type a plain-English question — get an instant answer with an auto-generated chart. Save the question, or act on the result: turn it into cases, or into an email to exactly those students. No report-writers, no SQL.

Early warning that explains itself.

At-risk scores across attendance, grades, engagement and wellbeing are transparent and explainable — you can see exactly why a student was flagged. Each flag is one click from a real pastoral-care action, not just a number on a dashboard.

How many international students are behind on fees this term?
14 students · $38,410 outstanding
Save question Email these students Create cases
Why was Tema flagged?
risk 72
{{ r.label }}
{{ r.v }}
→ Open a pastoral-care case

For teaching staff

The tedious parts of teaching, taken off the desk.

{{ s.tag }}
{{ s.t }}

{{ s.b }}

Governance

The AI Control Centre. Not a black box.

You govern your own AI: manage agents, choose models, review approval-gated actions, set policies, test prompts in a sandbox, and watch usage analytics. AI in Pallara is administered like any other enterprise system.

{{ a.t }}. {{ a.b }}
AI Control Centre agents · models · approvals · policies
{{ ag.name }}
{{ ag.model }}
{{ ag.status }}
1 action awaiting approval — bulk email draft to 34 students. Review →
Privacy by design

Academic access is scoped to the authenticated learner, providers choose their model stack, and institution data is isolated at the database level.

Approval-gated actions

Agents propose; your staff approve. Anything consequential waits for a human.

Sovereignty-ready

Sensitive workloads can run on locally hosted models, in your own environment or region.

See Aria answer questions from your own course material.

Book a walkthrough Next: compliance →