Business intelligence is very good at telling you what happened, and almost silent about what happens next. A dashboard shows forty unmatched reconciliation items; someone still has to open forty tickets, chase forty owners and write forty notes. AI agents can take on that second half — if the handoff is designed so that people stay in charge. This guide explains how to connect the two, and how ATX AnalyticsTable and ATX Agent Hub divide the work.
A BI tool's job ends at the screen or the report. Whatever follows — triage, follow-up, correspondence, a change in another system — is manual, and that is where most of the time goes. Adding an AI chat window to the dashboard does not change this; it produces a better explanation of the same problem.
ATX AnalyticsTable finds. Repeatable pipelines prepare and join the data, anomaly and outlier detection runs over the full dataset, and alert rules watch a pipeline, a query or a range on their own schedule. The result is a precise, masked table of exceptions.
ATX Agent Hub acts. Its AI agents are workflows that read the exceptions, look up context in your knowledge bases, draft the follow-up — a Jira issue, an e-mail, a status update — and wait for a person to approve every consequential step.
There is no proprietary link between the two products; the handoff runs through interfaces both already support, which keeps each side replaceable:
Automating action is where AI risk becomes real. Human-in-the-loop approval nodes, AI Guardian and Domain Constraint checks, immutable workflow versions with four-eyes promotion and full trace logs are what let a bank or insurer put an agent behind a BI finding without handing it authority it should not have.
A reconciliation pipeline in ATX AnalyticsTable runs after the overnight batch and exports the unmatched items, with account holder names masked. An ATX Agent Hub workflow reads the file each morning, groups breaks by likely cause using the reconciliation procedures in its knowledge base, and drafts one Jira issue per group with the evidence attached. A team lead approves the drafts before they are created, and the dashboard shows the breaks closing as the issues are resolved.
Yes, through a handoff point both sides support. ATX AnalyticsTable can export the exceptions to object storage or a database table on a schedule, and an ATX Agent Hub workflow scheduled after it picks them up. There is no built-in one-click connection between the two products today.
Only what you let through. Masking runs as a pipeline step before the export, so the agent receives masked values, and every agent run is recorded in a trace log.
Only where you design it to. Approval nodes can be placed before any consequential step, and requesters cannot approve their own changes when a workflow is promoted to production.
This site needs JavaScript for the full experience. Every page above is reachable without it.