Voice
Ask about the school records you already have permission to see — by text, or by voice in 11 languages once the voice provider is set up.
Aira — AI, Governed by People.
One assistant across chat, voice and automation. Aira answers from permission-checked school data and prepares each action for a person to confirm; anything that runs on its own runs only under an automation your school switched on.
Aira · governed school intelligence
Ask Aira AI. Get grounded answers.
Choose a school workflow. Aira uses tenant- and permission-checked product context, and separates read-only answers from action proposals that require confirmation.
Fee context Attendance review Enquiry follow-up Task inbox
School-controlled boundary
Your role, enabled modules and configured providers remain the operating boundary. Aira does not become the authority.
Aira AI
Grounded sample ready
Prepare an overdue-fee reminder for Class XI.
I can prepare a current preview of eligible overdue invoices and guardian recipients, including the recipient count, invoice count and total due. Review the preview, then approve or cancel.
Fee ledgerPermission scopeRecovery safeguard
Nothing has been sent.
Human checkpoint
Approval required for a supported send
This guided public preview recognises fee, attendance, admissions and assigned-task examples. It does not query live school data.
School-workflow sample context the live assistant uses public product and Help Center content
- Role-scoped contextExisting permissions shape what Aira can retrieve.
- Human approval gatesMaterial proposals wait for an authorized person.
- Action evidenceSupported executions retain acting-user and outcome context.
- 11 language optionsEnglish plus 10 Indian languages where configured.
Interactive governance path
Follow one request to the human checkpoint.
Choose any stage, or let the short guided sequence advance while the section is visible. It stops before execution and waits for an explicit sample approval.
Aira governance path
Illustrative workflow · sample data
Stage 01 / 06Review gate ahead
Sample operational request
“Prepare an overdue-fee reminder for Class XI.”
01 · Intent
Start with a bounded school task
Aira identifies the requested outcome and keeps the task inside supported ERP workflows.
Control evidence
Sample request: prepare a fee reminder for a selected class.
This control advances a sample only. It does not contact a parent or alter a live record.
Deployment boundary
Availability depends on the enabled module, user permission and configured channel. Drafted communication remains reviewable before a supported send action.
Interactive samples illustrate the control model; they do not read live tenant data or execute a live action.
The Aira AI & Automation Suite
One AI. Every corner of the school.
Aira works inside fees, compliance, results and the front office — answering from records the asking person is already allowed to see, and proposing actions a person confirms.
Governed AI capabilities 11 languages, text & voice
- 1AskText, voice or a product workflow
- 2ProtectPermissions checked; configured safeguards applied
- 3ComputeProduct engines calculate the real answer
- 4ReviewA person confirms every sensitive action
Example runIllustrative run · scored overdue accounts and prepared a human-reviewed reminder queue
6 AI copilotssupported figures computed by product engines— tap any card for how it works
Agent
Scores eligible overdue accounts, prepares a five-step recovery ladder and checks invoice state before each configured reminder.
Role-limited
Pre-submission DPDP, UDISE+ and APAAR scans with plain-language fixes and rule citations — before the deadline finds you.
Role-limited
“How many Class XI students have pending fees?” Plain-English answers from live school data — the model picks a report from a fixed list, it never writes the query.
Auto
Attendance, at-risk students, fee collections and open tasks, distilled into one daily “what needs you today”.
Role-limited
Duplicate-payment and outlier alerts, a weekly cash-flow forecast, bank-to-ledger reconciliation — and “ask about your books” in plain English.
- Common identifier patterns are redacted by default; supplied known names can be tokenised before supported AI calls.
- Supported figures are computed by product engines; AI provides the explanatory text.
- You control it: per-persona on/off toggles, credit budgets, and an off-switch.
The questions schools ask about AI.
What data can Aira see?
Aira checks the signed-in user, the active school and the relevant permissions before any supported source is used, so it answers only from permission-checked product context for that school. Parent Aira is designed to use the authenticated family's child scope, and the public marketing assistant grounds only on public content.
What happens when the AI is wrong?
Model-written text can be wrong, so Aira separates read-only answers from action proposals. Nothing material runs on Aira's own authority: a person reviews and confirms each supported action, the acting user and outcome are recorded, and material answers should be verified before you rely on them.
Which numbers are deterministic?
Operational numbers — fee totals, attendance counts, dues and similar signals — are computed from recorded school data and configured rules, not generated by a language model. Aira explains those deterministic results; the model writes the prose around them.
How does per-plan availability work?
Aira features follow your plan and stay off until you enable them. Availability also depends on the enabled module, the user's permission and any configured channel, and the pricing page's live feature comparison shows which plans include the AI suite.
See Aira on your school's own workflows
Start the 90-day trial — no card — or walk one workflow with us first. Aira features follow your plan and stay off until you enable them.