How to Automate Approval Workflows (With Templates)
TL;DR:
- Approval workflows are the most common first automation project because they’re high-frequency, rule-based, and universally frustrating when manual
- Three patterns cover most use cases: sequential (one approver after another), parallel (multiple approvers simultaneously), and conditional (different approvers based on criteria)
- The biggest mistake is replicating unnecessary approval steps; automate the routing, but first eliminate approvals that add time without reducing risk
- Every automated approval needs three things: an escalation path for non-responsive approvers, an exception path for edge cases, and an audit trail for compliance
Approval workflows are where most organizations start automating, and for good reason. The manual version is universally painful: someone submits a request, it sits in an approver’s inbox for days, nobody knows the status, and when the approval finally comes through, the need has often changed. Automating approvals eliminates the waiting, standardizes the routing, and produces the documentation that compliance teams require.
About 60% of businesses have implemented automation in at least one workflow, and approval routing is consistently the first or second use case. The technology is straightforward. The harder part is designing the approval logic correctly: which requests need approval, from whom, under what conditions, and what happens when the process breaks down.
For real-world examples of approval workflows across departments, see our 20 workflow automation use cases. For the implementation methodology, see our step-by-step playbook.
Three Approval Patterns
Sequential Approval
The request moves through approvers one at a time in a defined order. Approver A reviews first. If approved, it moves to Approver B. If Approver B approves, it moves to Approver C.
Use when: The process requires incremental sign-off (department head before VP before CFO) and each approver’s decision depends on the previous approver’s judgment.
Template: Employee submits expense report. System validates against policy (auto-reject obvious violations). Routes to direct manager. If approved, routes to finance for budget confirmation. If approved, queues for reimbursement. Each step has a 48-hour SLA with automated reminder at 24 hours and escalation at 48.
Parallel Approval
The request goes to multiple approvers simultaneously. All must approve (unanimous) or a defined majority must approve (quorum) before the request proceeds.
Use when: Multiple stakeholders need to weigh in independently and their evaluations don’t depend on each other. Legal reviews the contract terms while finance reviews the budget impact while the hiring manager reviews the candidate qualifications.
Template: New vendor request submitted with documentation. System sends simultaneously to legal (contract review), finance (credit check), and operations (capability assessment). All three must approve. If any rejects, the request returns to the submitter with the rejection reason from each reviewer. If all approve within 5 business days, the vendor is onboarded.
Conditional Approval
Different approvers or approval paths activate based on request attributes. If the purchase is under $1,000, auto-approve. Between $1,000 and $10,000, route to department head. Over $10,000, route to VP. Over $50,000, route to VP then CFO.
Use when: Approval requirements scale with risk, value, or complexity. Most mature approval workflows use conditional logic because not every request warrants the same level of oversight.
Template: Purchase requisition submitted with amount, category, and business justification. System evaluates amount against approval thresholds. Under $500: auto-approve with notification to manager. $500 to $5,000: manager approval required. $5,000 to $25,000: manager then director. Over $25,000: manager, director, then VP. Each tier has SLA and escalation rules.
Designing Your Approval Workflow
Step 1: Eliminate Unnecessary Approvals
Before automating, question every approval step. Each one adds cycle time. Each one is justified only if it genuinely reduces risk. Ask: “In the last 100 requests, how many times did this approver reject or significantly modify the request?” If the answer is fewer than 5%, the approval step is adding time without reducing risk. Consider replacing it with a notification (the approver sees the request and can intervene if needed, but the process doesn’t wait for their response).
This isn’t a governance shortcut. It’s governance discipline. Seampoint’s research on 18,898 tasks shows that governance constraints (consequence of error, verification cost, accountability requirements) should determine oversight levels. An approval step exists to catch errors whose consequences justify the delay. If the consequences don’t justify the delay, the step is bureaucratic friction, not governance.
Step 2: Define Escalation Rules
Every automated approval needs an answer to: “What happens when the approver doesn’t respond?” Without escalation rules, automated approvals create the same bottleneck as manual ones, just with better tracking of the delay.
Standard escalation pattern: send the initial approval request with a defined SLA (e.g., 48 business hours). Send a reminder at 50% of the SLA (24 hours). If the SLA expires, escalate to the approver’s manager with the original request, the approval history, and a note that the original approver’s SLA has expired. If the escalation target also doesn’t respond within 24 hours, route to a designated backup approver.
Step 3: Design Exception Paths
Approval workflows handle the 80% of cases that fit standard patterns. The remaining 20% (urgent requests that can’t wait for normal routing, requests that don’t fit any defined category, approvers who are on extended leave) need explicit exception paths.
Common exception patterns: an “urgent” flag that bypasses the standard SLA and sends to all approvers simultaneously with a shorter response window. An “out of office” detector that automatically reroutes to the designated delegate. A “doesn’t fit categories” path that routes to a human coordinator who determines the appropriate approval chain manually.
Step 4: Build the Audit Trail
Every automated approval should log: who submitted the request (with timestamp), who received the approval notification (with timestamp), who approved or rejected (with timestamp and any comments), and what the final disposition was. This audit trail is produced automatically by most workflow automation platforms and satisfies compliance requirements for documented approval processes.
Building in Common Platforms
In Zapier: Use a form submission (Google Forms, Typeform, or Zapier Forms) as the trigger. Add a “Send Email” or “Send Slack Message” step for the approval notification. Use Zapier’s “Delay” and “Filter” steps to implement waiting periods and conditional routing. For sophisticated approval routing, Zapier’s Paths feature handles conditional logic.
In Make: Build the workflow as a visual scenario with router modules for conditional paths. Use the “Sleep” module for delays and the “Email/Slack” modules for notifications. Make’s error routes handle failed notification delivery. The visual canvas makes complex multi-branch approval workflows easier to design and maintain than Zapier’s linear structure.
In Power Automate: Use the built-in “Start and wait for an approval” action, which provides a native approval interface within Teams and Outlook. Power Automate’s approval actions are the most polished among the three platforms for this specific use case, with built-in response tracking, mobile approval support, and integration with Microsoft 365.
For platform selection guidance, see our Zapier vs. Make vs. Power Automate comparison. For the broader tools landscape, see our workflow automation tools guide.
Common Approval Automation Mistakes
Replicating the manual process exactly. The manual approval process probably includes steps that exist because “we’ve always done it that way.” Automating these steps preserves the waste at machine speed. Redesign before automating.
No escalation for non-responsive approvers. An automated approval without escalation rules creates a polite bottleneck: the request waits patiently instead of someone chasing it, but it still waits.
Over-automating the decision. Auto-approving based on rules works for low-risk, high-volume requests (office supplies under $100). It’s inappropriate for requests where the approver’s judgment genuinely adds value (headcount additions, contract modifications, exceptions to policy). The routing should be automated. Whether the decision itself should be automated depends on the consequence of error.
No feedback to the submitter. Automated approvals should notify the submitter at every stage: request received, request sent to [approver name], approved/rejected with reason, and action taken. Silence during the process creates the same anxiety as the manual version, even if the process is actually moving faster.
Frequently Asked Questions
What is the best tool for automating approval workflows?
Power Automate has the most polished native approval features (built-in approval actions with Teams/Outlook integration). Zapier and Make are better when the approval workflow spans non-Microsoft applications. For simple approvals, all three work well. For complex conditional routing, Make’s visual builder is the most capable.
How do I handle approvers who are on vacation?
Build delegation rules into the workflow. When an approver’s out-of-office is active (detectable in most platforms through calendar integration), the workflow automatically routes to their designated delegate. If no delegate is configured, escalate to their manager.
Should I auto-approve low-value requests?
Yes, with appropriate thresholds based on your organization’s risk tolerance. Auto-approval for routine, low-value requests (office supplies, standard software subscriptions, pre-approved vendor reorders) eliminates unnecessary delays. Set the threshold based on the consequence of error: what’s the worst that happens if a request at this dollar amount is approved incorrectly?
How do I measure approval workflow performance?
Track four metrics: average approval cycle time (from submission to final decision), SLA compliance rate (percentage of approvals completed within the defined timeframe), rejection rate (percentage of requests rejected, indicating whether submissions are well-informed), and escalation rate (percentage requiring escalation, indicating whether primary approvers are responsive).