Why approval follow-ups matter more than the approval itself
Power Automate for approval follow-ups is really about fixing the part of the process that usually breaks: what happens after you click “Start and wait for an approval” and then… nothing.
That silence is where work gets stuck. A manager misses the Teams notification. An approver is out for two days. A finance request sits in Outlook under a pile of meeting invites. Suddenly your “automated” process still needs somebody to go chase people down by hand.
I’ve seen this a lot. Teams build the approval step first because it feels like the main event. But the part that makes the whole thing reliable is the follow-up logic around it. Reminders, escalations, reassignment paths, due dates, status visibility—those are the pieces that turn a basic flow into something the business can actually lean on.
So if you’re building approvals for purchase requests, document sign-off, access requests, invoice reviews, or HR tasks, the follow-up design deserves just as much attention as the approval action itself.
What “approval follow-ups” actually includes
When people hear follow-ups, they usually think “send a reminder email after two days.” Sure, that’s part of it. In practice, though, approval follow-ups are a bigger pattern.
They usually include:
- Reminder messages after a set amount of time
- Escalation to a backup approver or manager
- Status updates back to the requester
- Reassignment when the original approver is unavailable
- Cancellation or expiration rules
- Logging so you can see who was asked, who responded, and when
That last item gets missed all the time. But when somebody asks why a request took six days, you need more than “the flow ran successfully.” You want a clean trail showing the approval went out Tuesday, a reminder fired Thursday, it escalated Friday, and finally wrapped up Monday.
The biggest mistake: putting all the logic inside one waiting step
Power Automate makes it easy to use the Approvals connector and then wait for a response. For simple cases, that’s fine. Once you need more realistic follow-up behavior, though, that single blocking step starts to get in the way.
While the flow is sitting there waiting, your options narrow pretty quickly. You can work timeout behavior into some designs, but a lot of the time you end up with logic that’s tougher to test, tougher to maintain, and honestly tougher to explain to the next person who inherits it.
A cleaner approach is often to separate the approval request from the follow-up engine. Create the approval, store the approval metadata somewhere reliable—such as SharePoint or Dataverse—and let scheduled flows check for pending items and decide what happens next.
You get more control that way. And when requirements change—and they will—the process is a lot less fragile.
A practical architecture that works
If you want Power Automate for approval follow-ups to stay maintainable, think in terms of two or three small flows instead of one giant one.
1. Submission flow
This flow starts when a request is created. Maybe that’s a Microsoft Form, a SharePoint list item, a Power Apps submission, or a Dataverse row.
Its job is to create the approval, write key details to your tracking source, and notify the first approver. Keep it focused.
2. Follow-up flow
This is usually a scheduled cloud flow that runs every hour or every business day. It checks for approvals still marked pending, compares the current time to the due date or last reminder date, and then takes the next action.
That next action might be a Teams reminder, an email to the approver, an update back to the requester, or an escalation to somebody else.
3. Completion flow
In some environments, it helps to have a separate flow handle what happens after approval or rejection. Update the source record, notify the requester, trigger downstream tasks, and close out the tracking record.
Breaking it up this way gives you something a lot of teams don’t think about at first: operational clarity. When a user says, “Why didn’t I get a reminder?” you know exactly which flow to inspect.
Choose the right place to track approval state
Your follow-up logic is only as good as the data behind it. If approval state isn’t stored somewhere structured, you wind up reverse-engineering status from emails and approval history, and that gets messy fast.
For a lot of Microsoft 365 teams, SharePoint Lists are a practical middle ground. They’re easy to set up, they work well with Power Automate, and for many internal approvals, they’re plenty.
Dataverse is stronger when you need tighter security, cleaner relational data, or more app-like behavior. If you’re already using Power Apps in a serious way, it’s often the better long-term fit.
The fields I usually want are pretty simple:
- Request ID
- Requester
- Current approver
- Approval ID
- Status
- Submitted date
- Due date
- Last reminder date
- Escalation level
- Final outcome
You don’t need a giant schema. Just enough structure so the follow-up flow can make decisions without guessing.
Reminders should be helpful, not annoying
A reminder strategy can absolutely improve response time. It can also train people to ignore your automation if you overdo it.
Good reminders are specific. They tell the approver what’s waiting, when it was submitted, what the due date is, and what happens if no action is taken. A vague “Please review” message is easy to skip. A message like “Purchase request PR-1048 for Adobe Acrobat is due tomorrow and will escalate to Finance Ops if no response is recorded” gets attention.
Channel choice matters too. For some teams, Outlook works fine. For others, a Teams message is way more visible. In a few cases, using both makes sense—but only if the process matters enough to justify the extra noise.
If you want a broader pattern for this, my post on reminders and escalations goes deeper into timing, message design, and escalation rules.
Escalation rules need business logic, not just timers
Lots of flows escalate after a fixed number of hours. That’s a decent starting point. Real business processes usually need a little more judgment baked in.
For example, a contract review might escalate to legal leadership. A PTO request might go to the approver’s manager. An access request for Microsoft 365 admin roles might need to reroute to a security group instead of one person.
Here’s the part people often miss: escalation does not always mean “approve on behalf of the original approver.” Sometimes it means “take ownership of the decision.” Other times it means “help unblock the original approver.” Those are different workflows.
Your escalation model should match the risk in the process. A software license purchase and a production access request should not be treated the same way, even if both use the Approvals connector.
Routing and follow-ups belong together
Follow-ups work a whole lot better when routing is designed properly up front. If you’re sending approvals to the wrong person, no reminder strategy is going to save you.
Maybe the approver depends on department, cost center, region, or dollar amount. Maybe requests over a threshold need a second stage. Maybe an absence mailbox or delegated approver should be used during vacations. Those routing choices directly shape how follow-ups should work.
That’s why I’d treat routing and follow-up logic as one design problem, not two separate ones. If you need ideas there, this guide on approval workflow routing is a good companion piece.
Use business calendars, not just elapsed time
One angle people don’t always think about: a “48-hour reminder” sounds reasonable until the request lands Friday at 4:30 PM.
Now your Monday morning reminder may already feel late, or your Sunday escalation may feel ridiculous. That’s where business time matters. If the approval ties into operations, payroll, procurement, or onboarding, the difference between clock time and working time changes how the process feels to users.
Power Automate doesn’t magically solve business-calendar logic for you. You may need a custom table for holidays, business-hour checks, or another lightweight helper flow that calculates the next valid reminder window. It is extra work, sure. But it usually makes the automation feel much more thoughtful—and a lot less like a bot blasting messages on a timer.
Build for exceptions from the start
Approvals rarely fail because the happy path is difficult. They fail because people go on leave, org charts change, request data is incomplete, or someone rejects a request without enough detail for the requester to fix it.
That’s why exception handling shouldn’t be bolted on later. It belongs in version one.
A few examples worth planning for:
- The approver account is disabled
- The approver does not have a manager in Entra ID
- The request is withdrawn while approval is pending
- The same item gets resubmitted after rejection
- The approval record exists, but your tracking item failed to update
If those cases sound familiar, you’ll probably also want to read about handling approval exceptions and some broader approval flow patterns that help keep edge cases from taking over the design.
Real use cases where follow-ups pay off quickly
This pattern shows up all over Microsoft 365 environments.
Purchase approvals
A SharePoint-based request form creates an approval request for the department manager. If there’s no response after one business day, send a Teams reminder. After two business days, escalate to Finance. If approved, create a task for procurement.
Access requests
A Power Apps form submits a request for a shared mailbox, Teams team, or SaaS tool. The direct manager often gets the first approval. If they don’t respond, reroute to a backup approver pulled from a support-owned mapping list. That beats IT having to chase people manually.
Policy acknowledgments and sign-offs
Some organizations use approvals for document sign-off, especially when a team lead or compliance reviewer needs to confirm a step was completed. Follow-ups are the difference between “we sent it” and “we can prove it was actively managed.”
HR onboarding tasks
This is a great fit for staged reminders because delays have a clear downstream cost. Miss one approval and the laptop order, badge request, or system access setup gets pushed back. If you’re working in that space, here’s a good example of approval follow-ups in onboarding.
What to send in each follow-up message
You don’t need fancy message formatting to make follow-ups effective. You do need the right details.
A solid reminder usually includes the request title, requester name, submission date, due date, a direct link to the approval or source item, and a short note on what happens next. If the request has business impact, say that too. “Delays may affect start date readiness” is a lot more useful than generic urgency language.
And send the requester updates at meaningful points, not every time the system breathes. People usually want to know the approval was submitted, reminded after the delay, escalated if needed, and then completed or rejected. More than that starts to feel noisy.
Testing follow-up logic is different from testing the approval itself
A lot of teams test whether the approval can be sent and responded to. That’s only half the job.
You also want to test overdue behavior, reminder frequency, escalation paths, reassignment, cancellation, and duplicate prevention. What happens if the approver responds right after the reminder but before the escalation flow runs? What happens if two scheduled runs pick up the same item? What happens when somebody edits the source request while the approval is still pending?
Those aren’t weird edge cases. That’s normal production behavior.
In my experience, adding a handful of explicit status values—such as Pending, Reminder Sent, Escalated, Completed, Rejected, and Cancelled—makes testing much easier. Logging each follow-up action to a history list or notes field helps too, instead of relying on memory and run history alone.
Start simple, but leave room to grow
You don’t need to build a giant approval platform on day one. Start with one reminder, one escalation rule, and one place to track the state. That alone can remove a surprising amount of manual chasing.
But build it in a way that leaves room for change. Approver delegation, vacation coverage, SLA reporting, and multi-stage escalation usually show up later. If the first version is modular, you can add those without tearing the whole thing apart.
If you’re implementing Power Automate for approval follow-ups this week, the best next step is pretty simple: pick one approval process that already causes manual chasing, map the reminder and escalation rules in plain English, and build the tracking list before you touch the flow designer. That one move saves rework later.

