Four Things That Decide Whether an Exo Integration Holds Up
Exo is a mature system with a real database behind it. That is an advantage and a trap, and these four choices are where the difference shows.
Reading is safe, writing needs a supported door
Pulling stock levels, debtor details, price data and order status out of the Exo database with well written read only queries is routine, low risk work, and it is how most useful reporting and website availability gets done. Writing is a different matter. Creating a sales order or an invoice by inserting rows skips the business logic, the ledger postings, the document numbering and the audit trail, and the damage usually surfaces weeks later in a reconciliation nobody can explain. Writes go through the documented interface, every time, even when it is slower.
Pricing in Exo is a structure, not a field
Exo holds price through policies, groups and quantity breaks that can produce a different price for the same item depending on the debtor, the quantity and the date. A web store that asks for the price of a product without asking on whose behalf will publish something confidently wrong. Any serious B2B integration has to resolve price the way Exo would, per customer and per quantity, and cache the answer with a sensible expiry rather than guessing from a base price and a discount percentage.
Change detection has to be designed
Exo will not tell you when something changed. That means the integration needs a reliable way to find new and modified records: a last modified column where one exists, a watermark table you maintain, or a comparison against the previous run. This sounds like a detail and it is actually the difference between an integration that notices a price change within minutes and one that quietly misses edits made by a user who was not following the expected process.
Build for the ERP you might have next
A good share of Exo sites are weighing a move to a cloud platform, and MYOB has made its longer term investment direction fairly clear. That is not a reason to delay integration work, because the manual handling costs you money today. It is a reason to build with the mapping and the business rules in an integration layer rather than hard coded into the store or the warehouse tools, so a later migration replaces one side of the connection instead of all of it.