Solutions / AI Navigator
AI Navigator sits inside your UMCA tenant, draws evidence from live transaction data and files your analysts upload, and accelerates alert triage, cluster investigation, and SAR drafting without any data leaving your environment.
Analysts start a new chat from the UMCA back office. History is grouped by Today, Yesterday, Previous 7 Days, and Older, so prior investigations are one click away.
Describe the question in natural language or drop a file into the chat. Navigator understands fraud terminology, transaction fields, and your platform's data model.
Navigator reads UMCA transaction data and the uploaded file, selects the relevant records, and reasons over them inside your tenant. Nothing is sent to an external LLM store.
Receive a clear answer with the records and signals that back it up. Follow up with more questions in the same chat and context carries through.
Every conversation is saved and searchable. Analysts can jump back to an investigation from last week, continue the thread, and share findings with the case record. History is grouped by date for fast lookup.
Drop a CSV, Excel, or text file into the chat and ask questions about it directly. Useful for reviewing exports, counterparty lists, or evidence sent from other teams without leaving the platform.
Navigator queries UMCA transaction, rule, and case data directly. Answers are backed by the records themselves, not a generic model's recollection. The analyst can inspect the evidence behind every statement.
Access is granted through the same role model as the rest of the platform. Grant it to fraud analysts, keep it off for other roles, or restrict it per tenant. No separate identity to manage.
A low-latency mode optimised for quick lookups and short questions during live case review. Pair it with the full reasoning mode for deeper analysis when the question demands it.
Navigator runs across issuing, acquiring, AML, and merchant risk tenants. The same assistant works for every line of business, respecting the data boundaries of each.
Generic LLMs answer from training data. AI Navigator reads your actual UMCA transactions, rules, and case history for every question. Every claim is backed by evidence an analyst can click through to.
Read the detailed comparison: Grounded AI vs. Generic LLMs for Fraud Analysis.
Access is controlled by the same role-based permissions that gate the rest of UMCA. Your data stays in your tenant. Nothing is persisted to an external model provider's store.
AI Navigator is deployed with the governance posture Canadian and global regulators are now writing into model risk rules. Learn how it fits OSFI E-23 and emerging model risk expectations.
