Automating Purchase Order Processing Using Power Automate Desktop
Introduction
In real-world business scenarios, employees often spend a significant amount of time switching between multiple systems to complete repetitive tasks. One such scenario is processing purchase orders, where data must be collected from different applications and matched with external sources like Excel.
In this blog, I will walk you through a practical automation use case implemented using Power Automate Desktop, where a bot processes purchase orders by integrating a web application, another system for lookup, and an Excel file.
Problem Statement
A web application displays a list of Purchase Orders (POs). However, these POs are incomplete and require additional details such as:
Ship Date
Order Total
Buyer State
These details are available in another system called Procurement Anywhere.
Additionally, based on the State, we must assign the correct Agent, which is stored in an Excel file (Agent Territory Mapping).
Objective
Build an automation that:
Extracts Purchase Orders from the main website
Logs into Procurement Anywhere
Retrieves Ship Date, Order Total, and State
Reads an Excel file to map State → Agent
Assigns the correct agent
Fills the form and submits it automatically
End-to-End Flow
Open the main challenge website
Download the Agent Territory Excel file
Read Excel data into a DataTable
Log in to Procurement Anywhere
Extract all PO numbers
Loop through each PO
Fetch details from Procurement system
Map State with Excel to find Agent
Fill details in the form
Submit the form
Implementation in Power Automate Desktop
Let’s break down the automation step by step.
Note:
You can try this challenge here:
You will need to log in to access the challenge.
Use the Community Login option and sign up if you don’t have an account.
You can use your Gmail or any valid email to register.
Make sure you are logged in before running the Power Automate Desktop flow, otherwise the automation may fail.
Step 1: Launch the Website
The bot starts by launching the main web page where purchase orders are displayed.
Launch Edge → Open challenge URL
This page contains the list of POs that need to be processed.
Step 2: Download Excel File
The automation clicks on the Download Agent Territory Spreadsheet button.
Press Button → Download Excel
This file contains the mapping of:
State → Assigned Agent
Step 3: Read Excel Data
The downloaded Excel file is opened and read into a DataTable.
Launch Excel
Get Table Range
Read Cells → ExcelData
Example:
| State | Agent |
|---|---|
| CA | John |
| NY | Mike |
This data will be used later for mapping.
Step 4: Login to Procurement System
The bot opens a new tab and logs into the Procurement Anywhere system.
Populate Email
Populate Password
Click Sign In
Once logged in, the bot can search for PO details.
Step 5: Extract Purchase Orders
The bot extracts all PO numbers from the main page.
Extract List → PONumberfromWebsite
Example:
PO001
PO002
PO003
Step 6: Loop Through Each PO
A loop is used to process each PO one by one.
LOOP FOREACH CurrentItem IN PONumberfromWebsite
This is the core logic of the automation.
Step 7: Fetch PO Details
For each PO:
Enter PO number in search
Extract details from table
Search PO
Extract Table → DataFromWebPage
Extracted fields:
Ship Date
Order Total
State
Step 8: Fill Ship Date
The ship date is directly populated into the form.
Populate Ship Date → DataFromWebPage
Step 9: Process Order Total
The order total comes with a dollar symbol, so it must be cleaned.
Remove $
Convert Text → Number
Example:
$500 → 500
Then populate it into the form.
Step 10: Map State to Agent
This is a key step.
The bot compares the state from the website with Excel data.
Find State in ExcelData
Example:
State = CA → Agent = John
Step 11: Assign Agent
The correct agent is selected in the dropdown.
Set Dropdown Value → Agent
Step 12: Repeat for All POs
The loop continues until all purchase orders are processed.
Step 13: Submit the Form
After filling all rows:
Click Submit
Step 14: Cleanup
Finally:
Close Excel
Delete downloaded file
This ensures the system remains clean.
Key Concepts Demonstrated
This use case covers important RPA concepts:
Web Automation
Launch browser
Extract data
Populate fields
Excel Automation
Read structured data
Perform lookups
Data Mapping
- Match State → Agent
Looping
- Process multiple records dynamically
Data Transformation
- Convert text to numeric values
Real-World Relevance
This scenario is very common in industries like:
Supply Chain
Procurement
Finance
Operations
Instead of manual effort, automation can:
✔ Reduce errors ✔ Save time ✔ Improve efficiency ✔ Handle large volumes of data
Conclusion
In this automation, we built a bot that integrates multiple systems to process purchase orders efficiently. By combining web automation, Excel lookup, and decision-making logic, the bot eliminates repetitive manual work.
This is a great example of how Power Automate Desktop can be used to solve real-world business problems using RPA.
