What Feishu / DingTalk / WeCom are good for
- Org directory, chat, and calendar already in use
- Standard leave and expenses whose fields match the suite template
- Pushing to-dos to phones so fewer items are missed
- Docs and meetings that belong in the suite anyway
They solve “people are online and messages arrive.” Early-stage or highly standard companies can run high-frequency approvals in the suite first. Do not pretend you have no collaboration tool.
What you are buying in a custom back office
- Routing by amount, department, and contract type
- A permission matrix that follows the role, with export for audit
- A contract ledger with seals bound to numbers, renewals in to-dos
- A current knowledge base that links to and from tickets
You are not paying for another chat. You are paying to avoid rework: the structure the suite cannot change, built once as a work entry you can maintain.
The real line is whether you can change structure for your business. Suites change notices easily; they struggle with “no seal without a contract number.” When finance must follow voucher state, the group gets “do it this way for now.” WeCom only, no back office of your own: see Is WeCom enough, or do you still need your own back office?. A packaged flow with too many clicks sends people back to WeChat; see A packaged OA nobody uses is almost as bad as having none.
The usual stack: the suite owns people and reminders; the custom back office owns status, permissions, and ledgers; to-dos can write back to Feishu or WeCom. Approvals that only live as chat nods still break at scale; see What happens when every approval lives in WeChat. How to estimate cost by scope: see What an OA / management system costs, and how long it takes.
DaXi does not reskin a suite and call it custom. If the brief is simple, we will say phase one can stay in Feishu. If it is complex, the plan states why custom is required—usually seals, contracts, multi-org permissions, and reconciling with ops systems. To map current use, go to the Management systems service page and say which app approvals run in today. Why a management system at all: see Why companies need a management system.
If you can change master data, it is your back office
Do not start the suite-versus-custom choice on price. Ask three questions: can contract renewals become to-dos automatically; do permissions follow the role on a transfer; can amount-over-limit routing change without calling the vendor. If all three are hard, you rented a flow. Custom is not “more unique is better”—it is making the few truly different approvals into hard rules, and still notifying Feishu, DingTalk, or WeCom for the rest.
Org shape matters too. Subsidiaries need isolation while HQ needs rollups; a suite department tree is often not enough. When sales, delivery, and finance share contract fields, a generic approval is hard to split. Those two cases: Group and subsidiaries: isolate permissions, still report up and Sales, delivery, and finance share contracts: how to split the back office. Estimate cost on “what the suite cannot change,” not “buy another OA brand.” DaXi draws the notice layer and the work layer on one diagram so two to-do lists do not compete.