A financial services company is automating its trade reconciliation process. A desktop flow, running unattended, retrieves trade data from a mainframe terminal, processes it, and compares it against data from a modern REST API. The process must be resilient to intermittent network failures when calling the REST API. The desktop flow must attempt to call the API up to three times with a 30-second delay between attempts before marking a trade as requiring manual review. Which action provides the most efficient and robust way to implement this retry logic directly within the desktop flow?
Answer and explanation
Correct answer: D
The most efficient and built-in method for handling this specific scenario is to use the native retry policy available on many desktop flow actions, including 'Invoke web service'. Setting the number of retries to 2 means the action will run once initially, and if it fails, it will be retried up to two more times, for a total of three attempts. This is the intended, declarative way to handle transient errors without building complex and less maintainable loop or subflow structures.
Question 2
A solution architect is designing an RPA implementation for a large-scale invoice processing center. Over 50,000 invoices must be processed daily by a fleet of 20 virtual machines. The processing time for each invoice varies significantly. The primary goals are to maximize machine utilization, ensure invoices are processed in the order they are received (FIFO), and provide a central point for monitoring progress and handling exceptions. Which combination of Power Automate components is best suited to meet these requirements?
Answer and explanation
Correct answer: C
This architecture is the canonical design for large-scale, resilient RPA. Work Queues provide built-in FIFO processing, state management (e.g., Queued, Processing, Processed, Exception), and centralized monitoring. A Machine Group allows the fleet of 20 VMs to act as a single pool of resources, automatically picking up the next available item from the queue, which maximizes utilization and provides load balancing.
Question 3
Multiple answers
A developer needs to create a desktop flow that interacts with a legacy Java-based application. The standard UI element recorder fails to correctly identify individual controls within the application's main window. The business requires that the automation be able to click specific buttons and enter text into specific fields identified by their accessibility properties. What should the developer do to enable reliable interaction with the Java application's controls? (Select TWO)
Answer and explanation
Correct answers: B, E
The Java Access Bridge is a technology that exposes accessibility information from Java applications to assistive technologies, which Power Automate for desktop leverages for UI automation.
After installing the Java Access Bridge, you must explicitly enable the Java automation feature within Power Automate for desktop so it can properly hook into the Java application's UI tree.
Question 4
A Power Platform administrator is reviewing the company's Data Loss Prevention (DLP) policies. They have configured a policy that places the 'HTTP' connector in the 'Blocked' group. A developer reports that their existing cloud flow, which calls a desktop flow, is now failing. The desktop flow itself does not use any HTTP actions, but the parent cloud flow does. What is the most likely reason for the failure?
Answer and explanation
Correct answer: C
The action in a cloud flow used to invoke a desktop flow is itself built upon the HTTP connector infrastructure. When the HTTP connector is placed in the 'Blocked' group by a DLP policy, this action is disabled, preventing the cloud flow from triggering the desktop flow, even if the desktop flow itself contains no blocked actions.
Question 5
You are developing a solution that requires passing a complex object with nested properties from a parent cloud flow to a child cloud flow. The child flow must then modify one of the nested properties and return the entire updated object to the parent.
Which combination of actions is required to correctly implement this data exchange?
Answer and explanation
Correct answer: D
For passing complex objects to a child flow, the standard practice is to pass the object (often from a 'Compose' action) as a whole. The child flow's trigger receives it as a string, which must then be parsed using 'Parse JSON' to reconstruct the object. To return the modified object, the 'Respond to a PowerApp or flow' action is used, which can output complex JSON objects back to the calling parent flow.
Question 6
True or False: When sharing a desktop flow with another user as a co-owner, the co-owner can edit the flow but must use their own connections to any connectors used within the flow.
Answer and explanation
Correct answer: A
This statement is true. Co-ownership grants permissions to edit, manage, and view the run history of a desktop flow. However, connections are tied to a specific user's credentials and are not shared. The new co-owner must create and use their own connections for the flow to run successfully under their context.
Question 7
A developer is using the 'Extract data from web page' action in a desktop flow to scrape a table of product information. The web page uses dynamic IDs for its HTML elements, causing the selector to fail intermittently. The table structure, however, remains consistent. Which approach should the developer use to create a more resilient selector for the table?
Answer and explanation
Correct answer: B
When dealing with dynamic attributes like IDs that have a static prefix or suffix, the best practice is to edit the selector manually. By changing the default 'equals' (=) operator to a 'contains' (*=), 'starts with' (^=), or 'ends with' ($=) operator, you can build a selector that matches the consistent part of the attribute, making it resilient to the dynamic changes.
Question 8
Case Study: Global Shipping Inc.
Company Background: Global Shipping Inc. (GSI) is a logistics and supply chain management company that operates worldwide. They handle freight forwarding, warehousing, and last-mile delivery. GSI is undergoing a digital transformation initiative to automate manual processes, reduce operational costs, and improve data accuracy. They have heavily invested in the Microsoft Power Platform.
Current Situation: A key bottleneck is the processing of customs declaration forms, which arrive as PDF documents in a shared mailbox. Currently, a team of data entry clerks manually reads each PDF and enters the information into three separate systems: a legacy AS/400 terminal application for freight management, a modern web-based CRM for customer records, and a Microsoft Dataverse table for analytics.
Automation Requirements:
The solution must monitor the shared mailbox for incoming emails with PDF attachments.
It must extract key fields from the PDF forms, including Shipper Name, Consignee Address, Declared Value, and a list of line items.
The extracted data must be entered into the AS/400 terminal, the web CRM, and the Dataverse table.
The solution must run unattended on a pool of virtual machines to handle high volume during peak hours.
Credentials for the AS/400 system and web CRM must be managed securely and not be hard-coded in the automation scripts.
A business continuity plan is required. If the primary Azure region hosting the VMs experiences an outage, the processing should fail over to a secondary region with minimal data loss.
Technical Constraints:
The AS/400 application has no modern APIs.
The web CRM has a REST API, but direct integration is a future phase; for now, UI automation must be used.
The company has a strict DLP policy that blocks most social media connectors but allows the HTTP, Dataverse, and Office 365 Outlook connectors.
You are tasked with designing the high-level architecture for this automation. Which design best meets all the requirements?
Answer and explanation
Correct answer: C
This design correctly separates concerns and uses the best components for each task. The cloud flow handles ingestion and intelligent processing. Work Queues provide the required scalability, resilience, and queuing. A Machine Group manages the pool of VMs for unattended execution. Storing credentials in Azure Key Vault is the most secure method. This architecture also supports the business continuity requirement, as Work Queue data is persistent and can be processed by machines in a secondary region if the primary fails.
Question 9
A developer is building a custom connector to an internal legacy system that uses a proprietary, non-standard authentication scheme. The scheme requires generating a unique token by combining a static API key with the current timestamp, hashing the result with MD5, and passing it in a custom HTTP header named 'X-Auth-Token'. Which feature of custom connectors must be used to implement this logic?
Answer and explanation
Correct answer: C
Standard authentication types and policy templates cannot handle this type of dynamic, multi-step token generation. This scenario requires writing custom C# code for the connector. By overriding methods like 'ExecuteAsync', a developer can programmatically access the request, generate the dynamic token using timestamps and hashing, add the custom 'X-Auth-Token' header, and then let the request proceed.
Question 10
An RPA team manages automations across multiple environments (Development, Test, Production). To avoid manual configuration changes during deployment, they use environment variables in their cloud flows to store the email address for failure notifications. The solution is deployed as managed. After deploying a new version to Production, an administrator needs to update the notification email address. What is the correct procedure for updating the environment variable's value in the Production environment?
Answer and explanation
Correct answer: D
Environment variables are designed to be configurable post-deployment without modifying the managed solution itself. The 'Default Value' is part of the solution definition, but the 'Current Value' is an environment-specific override. The correct ALM practice is to edit the 'Current Value' directly in the target environment (Production), which creates an unmanaged layer for that specific component's value.
