Discuss with our experts
Live demo

AI-powered support for EML Payments

Built around the questions EML cardholders and program partners ask every day — activating a new card, checking a balance, disputing a charge, replacing a lost or stolen card, reloading funds, chasing a pending disbursement, and getting back into the app. The agent answers and actions: activates cards, raises disputes, orders replacements, resends payouts, all from chat.

Try the EML Payments AI Agent

Loaded with sample data for an EML cardholder and a program manager running a disbursement program.

Click any prompt to copy it.

Make up any email and card last-4 (e.g. jordan@example.com and 1199) — the agent rolls with whatever you give it.

AU phone — coming

What our virtual support agent can help with

Five topic-matched workflows trained on EML's product disclosure and customer-support content. The agent quotes back the change before any write action.

Card activation and balance

Looks up a cardholder's card, activates a newly issued card, and surfaces the current balance and recent loads. Handles both prepaid and reloadable products.

Transactions and declines

Surfaces a recent transaction the cardholder doesn't recognise, raises a dispute with a reference, and explains the up-to-45-day investigation timeline. Flags fraud-shaped cases for the urgent-handling path.

Lost or stolen card replacement

Blocks the compromised card, orders a replacement (5-7 business days), and confirms the existing balance carries over. The first turn is always to stop further fraud.

Reloads and disbursements

Loads funds onto a reloadable card, computes the new balance, and confirms the funds are available immediately. Handles program-partner disbursement queries — checking a pending payout and resending it with a fresh claim link.

Account and program access

Diagnoses locked-out cardholders, sends a secure reset link to the registered mobile and email (30-minute expiry), and hands off cleanly when the contact details on file are out of date.

Stays in scope

No personalised financial advice, no "should I move my savings here", no decisions on whether a dispute will succeed. STEER guardrails route anything that crosses the line out to a human — without dead-ending the cardholder.

How we built this

Wired up specifically for EML's cardholder and program-partner support flows.

1

Read your real product disclosure

Grounded eleven knowledge base articles in EML's official Reloadable Visa Prepaid Card PDS and customer support pages. No fees, rates, or timelines fabricated.

2

Wrote the workflows

Five topic-matched Response workflows — card activation/balance, transactions/declines, lost-or-stolen, reloads/disbursements, account access — under one router that recognises cardholder vs program-partner perspective.

3

Mocked your tools

Twelve mock tools as six read-and-action pairs — getCard + activateCard, getTransaction + disputeTransaction, getCardStatus + replaceCard, getBalance + reloadCard, getDisbursement + resendDisbursement, getAccessStatus + resetAccess.

4

Stress-tested it

Ten conversation scenarios plus fraud-urgency and no-personalised-advice guardrails — verified across both cardholder and program-partner perspectives with real test tickets.

Built for regulated payments

Lorikeet is built for businesses handling sensitive card and account data — privacy, data residency, and regulatory boundaries are non-negotiable.

Multi-region data residency

EU, US, and Australian deployments. Pick the data region that suits your card-scheme and program-partner obligations — your data stays where it should.

Phased API access

Start read-only (card lookups, transaction queries, balance checks), then turn on write-back actions (activations, disputes, replacements, reloads) when you're ready. The agent only sees and actions what you've explicitly authorised.

Fraud-aware by design

The fraud-urgency guardrail recognises unauthorised-charge language and gates the agent into the urgent-handling path (block first, dispute second) rather than a generic FAQ reply.

Want to see this on your real data?

We'd love to walk you through how this works in production and answer any questions.

Discuss with our experts