Sage Intacct and Expensify integration

What moves between them

The main flow is from Expensify into Sage Intacct. Approved expense reports are read from Expensify on a schedule you set, typically after each approval wave. Each report is mapped to a GL entry in Intacct: the report amount is allocated to a GL account determined by the expense category, and the cost center dimension is set from Expensify policy tags. Employee names and corporate card merchant data flow as reference to enrich the GL line description. Reports are read-only in Expensify, so ml-connector never writes expense data back.

How ml-connector handles it

ml-connector manages two separate polling schedules: one for Expensify (fetching reports by approval date and status) and one for Sage Intacct (reading GL accounts and dimensions to validate the target allocation). When an Expensify report is approved, ml-connector queries the full report, maps each line to the configured GL account and cost center tag, then submits an Intacct GL journal entry via the XML gateway. Because Intacct uses a single XML endpoint and requires session management, ml-connector maintains the session token and automatically refreshes it before the 50-minute expiry. Intacct's XML responses are parsed for application-level errors (errormessage tags in the XML body, not HTTP status codes), and retried operations are deduplicated via Intacct's uniqueid flag. Expense category to GL account mapping is configured per customer in the flow UI, and missing mappings fail safely with an audit entry rather than creating unallocated entries.

A real-world example

A mid-sized professional services firm uses Sage Intacct for accounting and Expensify for employee expense reporting and corporate card management. Before the integration, the finance team reviewed approved expense reports in Expensify weekly, manually looked up the correct GL account and cost center for each project, then keyed the summary into Intacct's GL journal entry. This process was error-prone, created duplicate handling, and delayed month-end close. With Expensify and Sage Intacct connected, each approved report automatically posts to the correct GL account and project cost center. The team now reviews the GL entries in Intacct at month-end with expense reports already reconciled, eliminating the re-keying step and speeding close by two days.

What you can do

• •Post approved Expensify expense reports as GL journal entries into Sage Intacct on a configurable schedule.
• •Map Expensify expense categories and policy tags to Sage Intacct GL accounts and dimensions without manual lookup.
• •Maintain Sage Intacct XML gateway sessions and automatically refresh before expiry to prevent outages.
• •Handle Sage Intacct's application-level XML error responses and deduplicate retried entries via uniqueid.
• •Track expense report status and audit all GL postings with full transaction history.

Questions

Which direction does data move between Sage Intacct and Expensify?

The main flow is Expensify into Sage Intacct. Approved expense reports move from Expensify into Intacct as GL journal entries, allocated to the correct GL accounts and cost centers. Reference data such as GL accounts and dimensions are read from Intacct to validate the target allocation. Expense data is not written back to Expensify.

How does ml-connector handle Sage Intacct's XML gateway and session tokens?

ml-connector sends all requests to Intacct's single XML gateway endpoint at https://api.intacct.com/ia/xml/xmlgw.phtml. It manages the session token obtained from the initial getAPISession call, caches it for reuse, and automatically refreshes it before the 50-minute expiry window. Application-level errors are parsed from the XML response body, not HTTP status codes, so ml-connector checks for errormessage tags in every response.

What mapping is required between Expensify and Sage Intacct?

Each Expensify expense category must be mapped to a Sage Intacct GL account. Optionally, Expensify policy tags can be mapped to Intacct dimensions such as cost center or project. Missing mappings fail safely with an audit entry, and the mapping configuration is stored per customer in the flow settings.