TL;DR
Build high-volume approval and document flows first, since these Power Automate examples pay back fastest. The 16 examples below cover Microsoft 365, finance, HR, sales, service and IT, each with a trigger, setting and KPI. Standard-connector flows run on Microsoft 365 licenses, while premium connectors need Power Automate Premium at $15 per user per month. Before you build, score each idea on volume, stable rules, exceptions, system access and measurable value. Then give every flow error handling, a named owner and a data policy before real work depends on it. Use desktop flows only for apps with no API, and send bulk data or code-heavy work to Fabric pipelines or Logic Apps.
Key Takeaways
Start with high-volume, rule-based work such as approvals, invoice routing and email-to-SharePoint filing, where minutes saved are easy to count.
Every flow has five parts: a trigger, connectors, conditions, actions and error handling, and the last one decides whether it survives production.
Cloud flows run automatically, instantly or on a schedule, while desktop flows drive the screens of legacy apps that have no API.
Power Automate Premium costs $15 per user per month and a Process license $150 per bot per month, both paid yearly.
A flow run ends after 30 days and pending approvals time out with it, so long processes need reminders and escalation built in.
Governance means separate environments, data policies that block risky connectors, a named owner per flow and solutions for deployment.
Five Hundred Invoices, Six Touches Each
Take a finance team that receives 500 supplier invoices a month. Each invoice needs six manual touches: download it, check it, key it into the ERP, get it approved, file it and chase the late ones. That is 3,000 small tasks every month, and almost none of them needs human judgment.
Power Automate can take over most of those touches. A flow saves the attachment, reads the fields with AI Builder, routes the approval and files the result. People step in only for the exceptions, such as a missing PO number or a total that does not match.
The harder questions are which flow to build first and how to keep it running once people depend on it.
Why Power Automate Examples Beat Feature Lists
Feature lists tell you what a tool can do, while examples show what it does on a normal Tuesday. The same holds for enterprise workflow automation solutions in general. That difference matters because a working flow has to survive messy inputs, absent approvers and expired passwords.
The Power Automate examples in this post are written to be copied. Each one names its trigger, the setting that makes it reliable and the KPI that proves it worked. Some read as short scenarios, others carry a code snippet or a checklist, and one is built step by step later on.
Three questions frame everything that follows. Which flow should you pick, given your volumes and systems? How do you build it so failures are visible instead of silent? And how do you govern dozens of flows once teams start copying each other?
The business case is real, although it needs testing against your own numbers. Forrester’s Total Economic Impact study of Power Automate , commissioned by Microsoft and published in July 2024, modeled one composite organization. That composite had 30,000 employees and $10 billion in revenue. Over three years, the study reported a 248% return on investment and a net present value of $39.85 million.
Treat those figures as a modeled ceiling rather than a forecast. Your own baseline, measured before the first build, is the number that counts.
How a Power Automate Workflow Works
A Power Automate workflow, called a flow, is a set of steps that runs when something happens. For instance, a new email, a form response, a new table row or a clock time can all start one. Once you can name a flow’s parts, every example below becomes easier to read and to adapt.
The Five Parts of Every Flow
Every flow, from a two-step email alert to a multi-stage approval, is built from the same five parts. Knowing them also helps you spot what a template leaves out.
Trigger. The event that starts the run, such as “When a new email arrives” or a Recurrence schedule.Connectors. Prebuilt links to apps such as Outlook, SharePoint, Dynamics 365 or Salesforce, each exposing its own triggers and actions.Conditions. The rules that branch the flow, for example an amount over a threshold or a file type check.Actions. The work itself: create a record, post a message, start an approval or write a file.Error handling. Run-after settings and scopes that catch failures so that an owner hears about them.
Microsoft’s connector reference lists a large catalog of prebuilt connectors, split into standard and premium tiers. That split drives cost, as the licensing section shows, and our guide to Power Automate connectors covers the catalog in depth.
Error handling is still the part templates skip most often. Yet it decides whether a flow survives its first bad input, so every example here names the failure path as well as the happy path.
Cloud Flows, Desktop Flows and Generative Actions
Power Automate has more than one kind of flow, and picking the wrong kind is a common early mistake. Microsoft’s flow types documentation says cloud flows can run automatically, instantly or on a schedule. A scheduled flow might run every Monday at 9:00 AM. Desktop flows, by contrast, automate tasks on the web or the desktop through the screen, which makes them the answer for systems with no API.
Flow type How it starts Best for Example
Automated cloud flow An event, such as a new email or SharePoint item Reacting to business events File invoice attachments in a library Instant cloud flow A button, a Teams message or a Power Apps call Tasks a person starts on demand Request a laptop from a phone Scheduled cloud flow A Recurrence trigger Reports, reminders and clean-up jobs Weekly overdue-task digest Desktop flow Called from a cloud flow or run by a person Legacy apps with no API Key invoices into an old ERP screen Generative actions (preview) A step inside a cloud flow Tasks where you state the intent and AI picks the actions Early pilots only
Generative actions let you state only the intent of a step, and the AI then chooses the actions and their order. They are still in preview, so keep them out of anything finance or compliance depends on. If your team also builds apps, our comparison of Power Apps versus Power Automate shows where each tool fits. This overview of low-code automation platforms also places both in the wider market.
Copilot and AI in These Flows
You no longer have to build every flow by hand. With Copilot in Power Automate , you describe a flow in natural language, and Copilot proposes the trigger and actions. It runs on Azure OpenAI and is optimized for English, so write prompts in English and review every step before you save.
Copilot speeds up the first draft, but the five parts still apply. Treat its draft like a template and test it against real inputs before anyone depends on it. For agents that go beyond a single flow, see our guide to Copilot Studio and these AI agent examples .
16 Power Automate Examples at a Glance
Here is the whole catalog on one page. Use it to find the two or three examples closest to your backlog, then jump to those sections.
What the License Column Tells You
The license column matters as much as the effort column. The four Microsoft 365 flows use only standard connectors, so they can start on licenses you already have. Lead routing (example 11) is only a medium-effort build, but the Salesforce and Dataverse connectors are premium, so it waits on Power Automate Premium. Desktop flows need Premium or Process licenses, while AI Builder steps draw on AI Builder or Copilot Credits.
So check the tier before you promise a launch date, because an unbudgeted Premium license can hold up an otherwise finished flow.
Where to Find Power Automate Templates
You rarely need to start from a blank canvas, since templates exist for many common flows. Microsoft’s template documentation explains how to search or browse templates by category, and several examples here have a close template match.
A template is only a starting point, though. It still needs an owner, error handling, named connections and a solution to live in before real work depends on it.
Microsoft 365 and Document Examples
Document work is where most teams build their first flow, because the connectors are standard and the pain is visible. These four examples run inside Outlook, SharePoint, Forms, Planner and Teams, so they need no extra license beyond Microsoft 365.
1. Save Email Attachments to SharePoint
A shared mailbox for supplier statements fills with PDFs that someone drags into SharePoint every afternoon. A flow can instead file each one the moment it arrives. Start with the Outlook trigger “When a new email arrives (V3)” and set both Only with Attachments and Include Attachments to Yes.
Inside an Apply to each loop, a condition makes sure only PDFs move, and a Create file action writes each file to a library, with the sender and date in the file name:
endsWith(toLower(items('Apply_to_each')?['name']), '.pdf')The KPI is simple: the share of attachments filed with no manual touch. Watch for signature images and logos, too, since a missing file-type check fills the library with them.
2. Document Approval and Archive by Metadata
Policy documents often sit in a Draft library with Department, Status and Review Date columns. The trigger “When a file is created or modified (properties only)” would fire on every edit. So add a trigger condition that runs only when the author sets Status to Submitted:
@equals(triggerOutputs()?['body/Status/Value'], 'Submitted')Once the department head approves, the flow moves the file to the Published library and sets the Review Date one year ahead. A rejection sends the reviewer’s comments back to the author instead. Track the average approval cycle in days, because that is the number document owners complain about.
Talk to Kanerika
Turn Your Document Work Into Reliable Flows
Kanerika reviews your Microsoft 365 document and approval processes, then shortlists the flows worth building first, with owners and KPIs agreed up front.
Schedule a Demo → Picture a facilities request form that asks for the issue, the building and the urgency. Before the flow, every response arrived as an email that someone retyped into a task board. Now a flow does the retyping.
The flow uses “When a new response is submitted”, then “Get response details”, then “Create a task” in Planner. It maps urgency to the task priority and the building to a bucket, so the board sorts itself. Finally, the flow emails the requester a link to the task, which cuts the “any update?” messages. The KPI is the time from request to first assignment.
4. Overdue-Task Reminders and a Weekly Digest
Reminder flows are scheduled rather than event-driven. A Recurrence trigger runs every weekday at 8:00 AM, and a SharePoint “Get items” action pulls only the late, open items with a filter query:
Status ne 'Done' and DueDate lt '@{utcNow('yyyy-MM-dd')}'Filtering at the source matters, because looping over every item wastes requests against your daily limit. Each owner then gets one email listing their late items, and on Mondays a second branch posts a digest to the team’s Teams channel. Measure overdue items per week, since the line should fall within a month.
All four of these flows deal with documents and messages that live in Microsoft 365. When the document itself holds the data, such as a scanned form or contract, our guide to document intelligence explains how extraction fits in.
Finance and Procurement Examples
Finance is the function with the clearest return, because every flow touches money, deadlines or audit evidence. These three examples move from simple routing to AI-based extraction, and each one has a cost or control angle worth knowing before you build.
5. Purchase Order Approval by Amount Threshold
A purchase request saved in a SharePoint list or a Dataverse table starts the flow. A condition checks the amount, so small requests go to the line manager while larger ones also pass through finance. Each approval has a timeout, and an unanswered request escalates instead of waiting.
This is the example built step by step later in the post, including the rules table and the time-saved math. The KPI is approval cycle time, plus the share of orders approved before the supplier’s deadline.
Kanerika Service
RPA Consulting and Intelligent Automation
Kanerika designs and runs finance automations, from invoice capture to approval routing, with error handling and monitoring built in from the first flow.
Explore RPA Services 6. Invoice Data Extraction with AI Builder
This is the flow behind the 500-invoice scenario described earlier. Invoices land in a mailbox or library. AI Builder’s invoice processing model then reads the supplier, invoice number, PO number and totals, a form of automated data extraction . A condition then compares the extracted total with the PO amount from the ERP.
Matches within tolerance go straight to approval. Mismatches go to an accounts payable clerk with the fields already filled, so the clerk checks rather than types. The KPI is the touchless rate, that is, the share of invoices processed with no human edit.
Check the licensing before you design an extraction-heavy flow. Microsoft’s AI Builder credit guidance says existing customers can renew the AI Builder credit add-on until November 1, 2026. New customers must buy Copilot Credits instead, and the AI Builder credits seeded in Premium licenses also end on November 1, 2026. For the wider process, see our posts on accounts payable automation as well as intelligent document processing .
7. Expense Request Approval with Policy Checks
Expense approvals go wrong when managers approve claims they never really read. A flow can apply the written policy first, so the manager sees only claims that pass or a clear reason why one did not. Typical checks include:
Receipts. A claim above the receipt threshold with no attachment goes back to the employee.Limits. A meal or hotel amount above the daily limit is flagged for finance review.Duplicates. A filter query on amount, date and merchant catches the same claim submitted twice.
Only clean claims then reach the manager’s approval, which keeps the approval step meaningful. The KPI is policy exceptions caught before payment. Where finance work still runs through desktop screens, our guide to RPA in finance covers the screen-level options.
HR Examples
HR processes cross more teams than almost any other function. A single hire or departure touches HR, IT, facilities, payroll and the manager. So the value of a flow comes from coordinating those hand-offs between teams.
8. Onboarding Across HR and IT
A new-hire record in the HR system starts the flow, which fans out to three teams at once. IT gets a request to create the account and assign licenses, while facilities gets an equipment task and the manager a first-week checklist. Later, on the start date, the flow posts a welcome message in the team’s Teams channel.
Use parallel branches so IT and facilities work at the same time, not one after the other. The KPI is the share of new hires with a working account and equipment on day one.
9. Offboarding with an Access Removal Checklist
Offboarding is where a missed step costs the most, because an active account can outlive its owner’s employment. The trigger is the leaving date in the HR record, and a daily scheduled check catches records entered late.
The flow then creates a checklist covering account disablement, license removal, mailbox handover, device return and app access. Each item has an owner and a due date. Anything still open 24 hours after the last day escalates to the IT manager. The KPI is hours from the last day to full access removal.
10. Leave Requests with Manager Approval and Calendar Sync
Employees submit leave through a form, a Teams message or a Power Apps screen. The flow starts an approval with the manager. Once approved, it writes an all-day event to the employee’s Outlook calendar and to a shared team calendar.
It also updates a leave balance list, so HR no longer reconciles spreadsheets at month end. Set the approval to time out after three days, then send the manager a reminder. Average days from request to decision is the KPI to watch. For a wider view of flows like these, see our guide to AI workflow automation .
Sales, Marketing and Customer Service Examples
Customer-facing flows are judged in minutes, not days. A lead that waits an hour goes cold, and a missed SLA costs goodwill. So these examples focus on speed and on catching problems before a customer notices them.
11. Lead Routing to the Right Owner
When a web lead arrives in Dynamics 365 or Salesforce, the flow picks it up through their connectors. A Switch on region and company size assigns the owner, and a Teams message alerts that rep within a minute. If the lead still has no activity after 30 minutes, the flow reassigns it to the queue manager.
Plan the license before the build. The Salesforce and Dataverse connectors are premium, so this flow needs Power Automate Premium. The KPI is the median minutes from lead creation to first contact.
12. Support Case Escalation Before an SLA Breach
A scheduled flow runs every 15 minutes and finds open cases whose SLA deadline falls within the next hour. Each case then gets a priority bump, a note on the record and a Teams alert to the team lead. That gives the team an hour to act instead of a breach report the next morning.
A useful variant handles customer feedback. For instance, survey responses with a low score open a follow-up case for the account owner. Track the SLA breach rate per month for the escalation flow, and time to follow-up for the feedback variant.
13. Scheduled Social and Campaign Posting with Approval
Marketing drafts posts in a SharePoint list with the copy, image link and target date. Two days before each date, a flow sends the draft for approval. Approved posts go out through the relevant connector or into the publishing tool’s queue, and rejected drafts return to the writer with the reviewer’s comment.
The KPI is posts published on schedule compared to posts planned. It sounds minor, but missed campaign dates often trace back to slow approvals.
Sales and pricing teams automate research as well as outreach. In one Kanerika project for a travel booking business, Power Automate bots collected competitor prices from booking sites. The bots then filled a single discovery database, so analysts no longer reviewed prices by hand.
Case Study
38% Faster Pricing Discovery with Power Automate
A connected-mobility client in travel booking replaced manual price checks with Power Automate bots, cutting pricing discovery time by 38% and booking processing time by 38%.
Read the Case Study → IT, Data and Operations Examples
IT and data teams use flows as both a front door and a watchdog. Requests come in through email and forms, while jobs, loads and dashboards need someone watching them overnight. These three examples cover both jobs, as well as the legacy systems that have no API at all.
14. IT Ticket Triage and Routing
Requests reach the IT mailbox in every shape, from “laptop broken” to a forwarded chain of twenty emails. The flow reads each message, classifies it with keywords or an AI Builder category model. Then it creates a ticket and assigns the right queue.
Password-reset requests also get an instant reply with the self-service link, so they never reach a technician. The KPI is the share of tickets routed to the right queue on the first try.
15. Data-Sync Failure Alerts and Scheduled KPI Reports
Data teams need to know about a failed load before the business does. Suppose a nightly job is loading results into a Snowflake data warehouse . When it fails, a flow posts the error to a Teams channel and opens a ticket for the on-call engineer. Snowflake publishes its own connector in Microsoft’s catalog, so the flow can also query a load-status table after each run.
Power BI also adds a second pattern. A Power BI data alert can trigger a flow when a number crosses a threshold. However, alerts work only on dashboard tiles pinned from report visuals, and only on gauges, KPIs and cards. So design the dashboard with that limit in mind.
A scheduled flow can also export a KPI summary every Monday and send it to managers who never open the report. Our comparison of Power Automate vs Power BI explains how the two tools divide this work. The KPI here is minutes from failure to alert.
16. Desktop Flow That Keys Data into a Legacy ERP
Some systems have no API, such as an old ERP screen or a supplier portal. A desktop flow handles them instead, clicking and typing the way a person would. A cloud flow collects the approved invoice data, then calls the desktop flow, which runs unattended on a machine licensed under the Process plan.
Screen automation breaks when a layout changes, so capture a screenshot on every failure and cap the retries. For more patterns, see these RPA use cases . Measure entries per hour and failures per thousand runs.
Checklist
Robotic Process Automation Checklist
Check process stability, system access, exception handling and ownership before you automate a legacy screen with a desktop flow.
Get the Checklist → Worked Example: Build a Purchase Approval Flow Step by Step
Purchase approval is a strong first build for many teams. Volume is steady, the rules are already written down, and every delay has a cost, such as a missed early-payment discount. So here is the full build, from business rules to the time-saved math.
Inputs and Business Rules
Write the rules down before you open the designer. Finance should sign off on this table, because the flow will enforce it exactly as written.
Amount band Approvers Timeout Escalation
Under $5,000 Line manager 2 days Manager’s manager $5,000 to $25,000 Line manager, then finance controller 3 days Head of finance Over $25,000 Line manager, finance controller, then CFO 5 days CFO’s office, flagged in Teams
These thresholds are illustrative, so use your own delegation-of-authority policy. Also keep the amounts in a settings list so finance can change them without editing the flow.
Build It in Eight Steps
The build uses only standard actions. The order matters, though, because the error handling wraps around the approvals rather than sitting after them.
Create a SharePoint list called Purchase Requests with Amount, Supplier, Cost Center, Requester and Status columns.
Add the trigger “When an item is created” on that list.
Add a Condition that checks whether Amount is 5,000 or more, with a nested Condition for amounts over 25,000.
In each branch, add “Start and wait for an approval” for the approvers in the table. Then set its timeout in Settings, such as P2D for two days.
Wrap all the approval steps in a Scope named Try, then add a second Scope named Catch below it.
Next, on the Catch scope, open Configure run after and select “has failed” and “has timed out”.
Inside Catch, email the escalation contact and then set Status to Needs Attention.
After the approvals, set Status to Approved or Rejected and notify the requester with the approver’s comments.
In expression form, the top-band check in step 3 reads like this:
@greater(triggerOutputs()?['body/Amount'], 25000)Microsoft’s error-handling guidance describes the four run-after states: failed, timed out, skipped and succeeded. With the Try and Catch pattern, a timeout reaches a person instead of disappearing into run history. Approvals also appear in Teams and Outlook, so approvers do not need to learn a new app.
Test Before Production
One clean run in the designer proves only the happy path. Run these five cases before anyone relies on it:
Normal path. One request in each amount band reaches the right approvers in order.Missing data. A request with no cost center lands in Catch instead of failing silently.Duplicates. The same request saved twice still starts only one approval.Wrong approver. A requester with no manager on file escalates instead of stalling.Failure path. A timed-out approval emails the escalation contact.
Also remember the platform ceiling. Microsoft’s limits documentation says a run lasts at most 30 days, including pending approvals. Any approval that could wait that long needs reminders and escalation well before day 30.
Measure the Time Saved
Use one formula: monthly volume times minutes saved per item, converted to hours, times the loaded hourly rate, minus license cost. Here is an illustrative case with every assumption stated:
Volume: 400 purchase requests a month.
Time saved: 15 minutes per request in chasing, keying and filing.
Loaded labor rate: $50 an hour.
Licenses: 20 Premium users at $15 a month, budgeted for a later version that reads the ERP through a premium connector.
That is 400 × 15 = 6,000 minutes, or 100 hours a month. At $50 an hour, the saving is $5,000 a month, or $60,000 a year. Licenses cost 20 × $15 × 12 = $3,600 a year, which leaves a net figure of about $56,400.
The estimate is only as good as its baseline. So time twenty real requests before you build, then compare after a month in production.
Which Power Automate Example Should You Build First?
The best first flow is the one with the most repeated, rule-bound work and the fewest surprises. Every team believes its own process is the most urgent, so a simple score keeps the choice honest.
Rate each candidate from 1 to 5 on the five criteria below, then add the scores. The criteria reward volume and predictability, because those are what make a flow pay back quickly.
Criterion Scores 5 when Scores 1 when
Volume Hundreds of items a month A handful a month Rule stability Rules are written down and rarely change Rules change every quarter Exception rate Few items need a person Most items need judgment System access Every system has a connector Only screen access, no API Measurable value Baseline minutes or errors are known No baseline exists
For example, take an illustrative accounts payable team with 800 supplier invoices a month, choosing between three Power Automate examples from this post:
Invoice routing with AI Builder: 5 + 4 + 3 + 4 + 5 = 21.Desktop flow keying invoices into a legacy ERP: 4 + 4 + 2 + 2 + 4 = 16.Scheduled campaign posting: 2 + 3 + 4 + 5 + 2 = 16.
Invoice routing wins clearly, so it ships first. The ERP keying flow follows once routing has proved its numbers, because it needs a Process license and more testing.
As a working rule, a score of 18 or more marks a strong first build, while anything under 12 should wait. Our guide to RPA for enterprise covers how to grow a backlog like this into a program. This look at the benefits of hyperautomation then shows where flows fit in a wider automation plan.
Talk to Kanerika
Score Your Automation Backlog With Kanerika
Bring your candidate processes, and Kanerika will score them on volume, rules, exceptions, system access and value, then plan your first three flows.
Schedule a Demo → Licensing, Governance and When to Choose Another Tool
One flow is easy to own. Fifty flows across ten teams, however, need a cost model, known limits and clear rules for who can build what. They also need a clear line where flows stop being the right tool.
What Power Automate Costs in 2026
Power Automate pricing has a simple core with a few add-ons. The figures below are US list prices from Microsoft’s Power Automate pricing page , paid yearly. A 30-day trial is also available.
Plan US list price What it covers
Microsoft 365 apps Included Cloud flows with standard connectors only Power Automate Premium $15.00 per user per month Premium connectors, attended desktop flows, process and task mining, managed environments, Dataverse storage Power Automate Process $150.00 per bot per month Unattended desktop flows, or a flow licensed on its own Power Automate Hosted Process $215.00 per bot per month Unattended automation on a Microsoft-managed virtual machine on Azure Process Mining add-on $5,000 per tenant per month Process mining at tenant scale
Microsoft’s license types page adds that extra capacity comes through Process licenses or pay-as-you-go. For a review of what you already own across Microsoft products, see our guide to Microsoft licensing optimization .
Request Limits and Flow Time-Outs
Every flow draws on a daily request allowance tied to its owner’s license. Microsoft’s request limits page sets 40,000 requests per 24 hours for paid Power Platform user licenses such as Premium. Users of Microsoft 365 apps with Power Platform access, in contrast, get 6,000.
Limit Value
Requests per 24 hours, Premium and other paid user licenses 40,000 Requests per 24 hours, Microsoft 365 apps with Power Platform access 6,000 Maximum run duration, including pending approvals 30 days Run history retention 30 days Flow failing continuously is turned off after 14 days Flow with no trigger activity may be turned off after 90 days (not for Premium or Process owners) Recurrence interval 60 seconds minimum, 500 days maximum
Two of these limits catch teams out. A flow that fails for 14 days straight is switched off, so failure alerts must reach someone who acts on them. Long approvals must also finish inside the 30-day run window, or the pending step simply times out.
Governance: Environments, Data Policies and Ownership
Governance sounds like paperwork, but it is what lets fifty flows run without a crisis. In practice it rests on four controls.
The first is separate environments, so makers build in development, test in a test environment and deploy to production. Then come data policies. An admin can block a connector, and Microsoft’s data policy documentation notes that a maker then cannot even save a flow that uses it.
The third control is a named owner. Every production flow needs a person, plus a co-owner, who receives its failure alerts. The fourth is solutions, which package flows with their connections and environment variables so they move cleanly between environments.
Power Platform data policies are also separate from Microsoft Purview data loss prevention , which protects content across Microsoft 365. When a flow writes to a warehouse, give its service account the narrowest role it needs and follow your Snowflake security controls for that account.
When Not to Use Power Automate
Some jobs fit other tools better, and forcing them into a flow creates something fragile. Use this short table to route the work, then keep flows for what they do well.
Job Better fit Why
Code-heavy integration, B2B processes or Azure DevOps work Azure Logic Apps Built for advanced, codeful workflows Bulk data movement and transformation Fabric or Data Factory pipelines Designed for large data loads on a schedule Simple tasks between non-Microsoft SaaS apps Compare with Zapier Depends on your app stack Large RPA estates with complex orchestration Compare with UiPath Depends on bot count and skills
Microsoft’s comparison of Logic Apps and Power Automate calls both designer-first integration platforms. It positions Power Automate for business users and citizen developers, whereas Logic Apps suits advanced, codeful scenarios.
For tool choices outside Azure, our comparisons of Zapier vs Power Automate and UiPath vs Power Automate go deeper.
Teams that already run both tools often consolidate on one. The walkthrough below shows what a UiPath-to-Power Automate migration involves from end to end.
Watch on YouTube
UiPath to Power Automate Migration Made Simple | End-to-End Automation
A Kanerika walkthrough of migrating UiPath automations to Power Automate end to end, for teams consolidating two RPA tools into one.
Common Mistakes and Their Fixes
A short list of build habits causes the bulk of production failures. Each one is cheap to fix before go-live and expensive to fix after.
Mistake What breaks Fix
Flow owned by one person’s account The flow stops when that person leaves Service account plus co-owners No error handling Failures sit unseen in run history Try and Catch scopes with run-after alerts Trigger fires on every edit Duplicate approvals and wasted requests Trigger conditions Hard-coded emails and site URLs The flow breaks when moved to production Environment variables in a solution Looping over thousands of items Request limits and slow runs Filter queries at the source Approvals with no timeout Requests wait until the 30-day limit Timeouts and escalation paths
How Kanerika Builds Power Automate Workflows
Kanerika’s Power Automate work often starts with an existing automation estate rather than a blank page. Trax, a global leader in spend management for companies with complex supply chains, ran 16 business process automations on UiPath. When license fees kept rising, the company decided to move, and the renewal date set a hard 120-day deadline.
The codebase, however, was not simple. It had custom integrations across web, API, database, Excel, Word and Office 365 steps, built up over two years. So the first step was a rapid assessment of every automation, its dependencies and its custom integrations. That turned a vague migration into a list of known jobs.
From there, Kanerika’s FLIP RPA Migration Workbench automated the workflow mapping and migration, which cut the manual effort. A pilot phase then validated each workflow before the full go-live. Grayson Deitering of Trax put the outcome plainly: “They ensured a smooth transition with minimal downtime while optimizing our workflows for maximum efficiency.”
The results are on record in the Trax migration case study . The two-year UiPath codebase moved in 90 days, inside the 120-day window, with 50% less effort thanks to automated migration. Annual licensing costs also fell by 75%.
The same habits carry into new builds, like the pricing discovery bots described earlier in the sales section. Each flow gets a test against real inputs, a named owner and failure alerts before cutover. Governance is in place from day one rather than added after an incident.
Kanerika Service
UiPath to Power Automate Migration Services
Kanerika inventories your bots, rebuilds them on Power Automate connectors where APIs exist, and cuts over with owners, tests and failure alerts in place.
Explore Migration Services Start With One Flow You Can Measure
Sixteen examples give you a menu, and the next step is choosing from it. Start with one flow that has high volume, stable rules and a connector for every system it touches. Good candidates include invoice routing, purchase approvals and email-to-SharePoint filing.
Build it with the five parts in place, especially error handling, and measure the minutes before and after. Then add governance before the second and third flows arrive, because that is when ownership questions start.
Save desktop flows for systems that truly have no API. Bulk data and code-heavy work belong in pipelines or Logic Apps. If you want a second opinion on your shortlist, Kanerika can score it with you.
Of all these Power Automate examples, the one to build first is the high-volume, rule-based flow your team can measure within a month.
Frequently Asked Questions
What are the best Power Automate examples for business?
The best Power Automate examples for business are high-volume, rule-based flows: purchase and expense approvals, invoice routing, saving email attachments to SharePoint, onboarding hand-offs and SLA escalation alerts. They pay back fastest because the rules are stable, the connectors are standard or well supported, and the minutes saved per item are easy to measure against a baseline.
What are simple Power Automate examples for beginners?
Good beginner flows use one trigger and one or two actions. Try saving Outlook attachments to a SharePoint library, turning Microsoft Forms responses into Planner tasks, or sending a weekday reminder for overdue list items. These use standard connectors included with Microsoft 365, so you can practice triggers, conditions and filter queries without buying an extra license.
How long does it take to learn Power Automate?
Many business users can build a simple two or three step flow within a day, using templates and the designer. Approvals with conditions, error handling and filter queries usually take a few weeks of practice. Desktop flows, solutions and environment management take longer, and they are worth learning before flows reach production use.
What is the best way to start with Power Automate?
Pick one repetitive process with high volume, stable rules and a connector for every system involved. Measure the current minutes per item, build the flow with error handling and a named owner, and test the normal path, missing data, duplicates and failures. Compare results after a month, then use that evidence to choose the next flow.
Where can I find Power Automate templates?
Power Automate includes a template gallery you can search or browse by category, such as approvals, notifications and file handling. Open Templates in the left navigation of the Power Automate portal to find them. Treat a template as a starting point, then add error handling, a named owner and solution packaging before production use.
When should you not use Power Automate?
Avoid Power Automate for bulk data movement and transformation, which belongs in Fabric or Data Factory pipelines, and for code-heavy or B2B integrations, which suit Azure Logic Apps. It is also a poor fit for processes whose rules change constantly or need judgment on most items. Use it for repeatable, rule-based work instead.
What can Power Automate be used for?
Power Automate connects apps and automates repetitive work across Microsoft 365 and a large catalog of other services. Teams use it for approvals, notifications, document filing, data sync between systems, scheduled reports and alerts. Desktop flows extend it to legacy applications with no API by automating the screen, while AI Builder adds document extraction and text classification to flows.
What is the difference between a cloud flow and a desktop flow?
A cloud flow runs in Microsoft’s cloud and starts automatically from an event, instantly from a button, or on a schedule. It works through connectors and APIs. A desktop flow automates web or desktop applications through their user interface, clicking and typing like a person. Use desktop flows only when a system offers no API or connector.
How do you create an approval workflow in Power Automate?
Start with a trigger such as a new SharePoint list item, then add a Condition that routes by amount or type. Add Start and wait for an approval for each approver and set a timeout in its settings. Wrap the approvals in a Scope, add a Catch scope with Configure run after for failures, then notify the requester.
Can Power Automate process invoices automatically?
Yes. A flow can pick up invoices from a mailbox or library, extract supplier, invoice number and totals with AI Builder, compare them with the purchase order, and route matches for approval. Mismatches go to a person with fields filled in. AI Builder runs on AI Builder or Copilot Credits. Premium’s seeded credits end November 1, 2026.
Can Power Automate automate legacy ERP or desktop applications?
Yes, through desktop flows. A desktop flow automates an application’s screens by clicking, typing and reading fields, so it works with old ERP systems and portals that have no API. Attended desktop flows run under Power Automate Premium, while unattended runs need a Process license. Capture screenshots on failure, because layout changes break screen automation.
Does Power Automate cost extra beyond Microsoft 365?
Flows that use only standard connectors run on Microsoft 365 licenses at no extra cost. Premium connectors and attended desktop flows need Power Automate Premium at $15 per user per month, paid yearly. Unattended bots need the Process plan at $150 per bot per month, or Hosted Process at $215 per bot per month.
How do you monitor and fix failed Power Automate flows?
Check the flow’s run history, which keeps runs for 30 days, and open the failed action to read its error. Prevent silent failures by adding a Catch scope with Configure run after set to failed and timed out, then alert a named owner. Act quickly, because a flow failing continuously for 14 days is turned off.
Can Power Automate replace UiPath and other RPA tools?
It can for many workloads. Power Automate combines cloud flows for API-based work with desktop flows for screen automation, all inside the Microsoft ecosystem. Teams with large, complex RPA estates should compare bot counts, orchestration needs and skills first. Migrations work best when each automation is inventoried and rebuilt on connectors where an API exists.
