Explicit workflows
Configure approvals, notifications and escalation rules.
The challenges to address
Requests circulate by email and the same information is entered into several tools. Approvals stall without visibility and exceptions are handled outside the process.
Our response
We first clarify the intended workflow: triggers, rules, controls and exceptions. Automation targets a manageable scope, with human intervention where decisions require it.
What happens after an order arrives?
Choose an incident and move a fictional order forward. You control approval and recovery.
A controlled workflow checks references, waits for human approval and then transfers the order using the same key. A missing reference goes back to the sales owner; a duplicate keeps its existing status; an unavailable application waits for a controlled retry.
Enable JavaScript to explore this example.
- Fictional order key
DEMO-1042- New transfers
- 0
- ReceiptReceived
- Reference checksTo check
- Human approvalTo approve
- TransferNot transferred
Order received. Start the reference checks.
The same key follows this order through to transfer. It helps identify a request that has already been processed.
Fictional demonstration in your browser. No connected system, real order or saved data.
What the engagement can cover
Connected tools
Connect applications through Power Automate, Python scripts or APIs as appropriate.
Visible exceptions
Plan logs, recovery actions and ownership of error handling.
A map of rules and exceptions
Example rules for transferring an approved order, including explicit error handling.
| Trigger | Rule | Exception handling |
|---|---|---|
| Order approved | Check customer and product references | Block missing references |
| Transfer to ERP | Reuse a unique identifier | Check status before retrying |
| Response received | Record the outcome | Alert the owner if unsuccessful |
How the engagement works
Flow mapping, exception analysis, a prototype for a stable use case, failure and recovery testing, then phased rollout. A manual continuity procedure is defined with the team.
Prepare an initial scope.
A few practical points to make scoping and acceptance clearer.
What you can prepare
- One specific workflow, its trigger and the people involved at each stage.
- Anonymised examples of complete, incomplete, rejected and duplicate cases.
- The manual process to retain and the person responsible for failed cases.
What we need to agree together
- Which approvals must remain human even when the exchanges are automated?
- What should happen if an application stops responding during processing?
Test the normal path, then an incomplete case, a duplicate and an interruption. Check the final state, notification to the correct owner and recovery without duplicate processing; the team must also be able to find the manual procedure.