In this guide
A request arrives in a shared inbox. Someone copies it into a spreadsheet, asks who owns it and sends a reminder when nobody replies. The software is rarely the hard part. The hard part is deciding what should happen when the request is incomplete, duplicated or urgent.
These business process automation examples start with that decision. Each has a trigger, a routine action, a named owner and a place for exceptions. Use the patterns as starting points and adapt them to the way your team works.
Choose one repeated process before connecting five apps. You should be able to describe how it works on paper and measure its current delays or errors.
1. Route a new request to the right person
A form, email or support message creates a record. Simple rules use the request type, location or existing account to assign an owner and acknowledge receipt. The record gets a next action and a due date. This is useful when requests land in several inboxes and nobody can see which ones are still open.
Keep the message modest. It can confirm receipt and say when a person will respond; it should not promise a solution before anyone has read the details.
Trigger: New request received Routine action: Create or update one record, assign an owner, acknowledge receipt Exception: Missing contact details, duplicate record or unclear request type Human owner: Team member responsible for the response Measure: Unassigned requests, time to first useful response, reopened requests
If the request belongs to a sales pipeline, our separate CRM lead routing guide covers assignment and backup rules in more detail.
2. Put purchase requests through a clear approval
A staff member submits a purchase request with the supplier, amount, purpose and required date. A rule sends it to the relevant approver. Approval or rejection is recorded against the request, and the requester sees the decision. Microsoft Power Automate, for example, supports flows that wait for an approver response before continuing. [1]
The exception path matters more than the notification. Who covers an absent approver? What happens when the amount changes after approval? Do not treat an unanswered request as approved. Keep financial authority and payment controls with the people your business has assigned to them.
Trigger: Complete purchase request submitted Routine action: Route to the authorized approver and record the decision Exception: Missing information, changed amount or absent approver Human owner: Requester until a decision is recorded Measure: Requests awaiting approval, elapsed review time, rework after approval
Find the broken handoff in your revenue journey.
3. Handle appointment changes without losing the slot
A customer asks to change an appointment. The workflow can flag the existing booking, offer the change to a scheduler and send a confirmation only after the new time is actually reserved. A freed slot can be shown to a waitlist or booking team.
The risky version cancels the first booking as soon as a message arrives, then fails to reserve the replacement. Keep the current slot until the change is confirmed, and make conflicting requests visible to a person.
Trigger: Change request received Routine action: Create a scheduling task and hold the current booking Exception: No suitable slot, double booking or special preparation needed Human owner: Scheduler Measure: Unresolved changes, late cancellations, bookings changed without confirmation
4. Flag renewals that need a decision
A contract or subscription approaches its review date. The system sends the account owner a short brief: current terms, usage information that is actually available, open support issues and the decision deadline. The owner checks the record and chooses whether to contact the customer, revise terms or leave the arrangement as it is.
Automating a reminder is straightforward. Automating the commercial decision usually is not. Missing terms or disputed invoices should pause the routine path. Avoid sending a customer a renewal promise based on a stale CRM field.
Trigger: Review date is approaching Routine action: Prepare a review task and link the source agreement Exception: Missing agreement, billing dispute or changed customer contact Human owner: Account owner Measure: Reviews completed before deadline, missing records, avoidable last-minute renewals
5. Turn a weekly report into an exception queue
A weekly export compares records that should agree: orders and fulfilment, bookings and attendance, or CRM stage changes and completed handoffs. The workflow highlights mismatches and sends them to the right person. It does not quietly overwrite one system with another.
Define which system is authoritative for each field. Keep a record of what was compared, when it ran and which differences were resolved. Microsoft recommends checking a flow’s run history and errors when troubleshooting failed automation. A report that never ran should not look like a clean week. [2]
Trigger: Scheduled reconciliation Routine action: Compare records and create an exception list Exception: Missing export, changed file format or unmatched identifier Human owner: Operations or finance reviewer Measure: Unresolved mismatches, age of exceptions, failed runs
How to choose the first workflow
Pick a process that happens often enough to observe, has a clear start and end, and has an owner who can explain the exceptions. A small pilot is easier to review than a promise to automate the whole business. Start with real examples of messy inputs, not just the ideal case.
- Write down the trigger, the desired result and the systems involved.
- Count a recent sample: volume, active handling time, corrections and delays.
- Name the person who reviews exceptions and the backup when they are away.
- Check whether one source record can trigger the same action twice. Use an identifier or another duplicate check before creating records or sending messages.
- Test a complete case, an incomplete case, a duplicate and a failed connection before switching on the workflow.
Platform tools can connect triggers and actions, but connected apps and permissions still need testing. [3] Microsoft also documents trigger settings for overlapping runs, which matters when two updates reach the same record close together. [4]
First pilot: [one workflow] Current volume and observation period: [measured] Baseline delay or correction rate: [measured] Routine steps to automate: [specific actions] Exceptions and reviewer: [specific conditions and owner] Duplicate prevention: [identifier or rule] Failure alert and manual fallback: [person and process] Pilot review date: [date] Decision: [keep, revise or stop]
Use our automation ROI guide to estimate the work removed and the work that remains. If the process is still ambiguous, simplify it with the team first. A faster version of a confusing process is still confusing.
A useful business process automation project has a clear trigger, an accountable owner, an exception path and a measure you can review. Start with one workflow and test its failure cases before expanding.