A financial services firm is automating a loan origination process that interacts with a legacy mainframe application. The solution design specifies that a single, monolithic Visual Business Object (VBO) will handle all interactions with the mainframe, including login, navigating through multiple screens, data entry, and logout. The process works correctly during UAT with a single robot. However, during performance testing with five concurrent robots, frequent session termination errors occur on the mainframe. The mainframe has a hard limit of five concurrent user sessions. Which design modification represents the most robust and scalable solution to this problem?
Answer and explanation
Correct answer: B
This is the best practice for managing limited, resource-intensive connections like mainframe sessions. Separating session management from business logic allows for the creation of a session 'pool'. A central session manager object can control the five available sessions, handing them out to worker processes as needed and reclaiming them upon completion. This prevents contention, maximizes the use of available sessions, and scales effectively without serializing the entire process.
Question 2
Multiple answers
A process is designed to handle insurance claims. It pulls items from a work queue, processes them, and is expected to run 24/7. During a production run, a NullReferenceException occurs deep within a sub-process called 'Process_Claim_Details'. The current exception handling is configured at the Main Page level only, with a single Recovery block that marks the queue item as an exception and gets the next case. This design causes the robot to enter a continuous loop, repeatedly failing on the same case without resolution. Which two modifications should be made to create a more resilient and informative exception handling mechanism? (Select TWO)
Answer and explanation
Correct answers: A, D
Placing an exception handler closer to the source of the error allows for more specific details to be logged (e.g., the values of local data items at the time of failure). Re-throwing the exception (bubbling it up) allows the main process to handle the business-level recovery, such as marking the work queue item as an exception.
This design prevents infinite loops for persistent system exceptions. By tracking the number of attempts for a specific item, the process can distinguish between transient errors (which might be resolved by a retry) and permanent failures. After a set number of attempts, the item should be marked as a business exception for manual review.
Question 3
A Blue Prism process designed to scrape data from a web application is running significantly slower than expected. A solution designer is using the debugger to investigate. They have placed breakpoints and are stepping through the process. They observe that a 'Wait' stage, configured to wait for a specific element to exist on a web page, is consistently taking the full 30-second timeout period before proceeding, even though the element appears on the screen within 2-3 seconds. What is the most likely cause of this debugging observation?
Answer and explanation
Correct answer: C
This is a classic performance issue. If the attributes defined in the Application Modeller for the target element match multiple elements on the page, Blue Prism cannot resolve the ambiguity. It will continue to search for a unique match until the Wait stage's timeout is reached. The element might be visible, but Blue Prism cannot confirm it is the correct element. The solution is to refine the attributes in the Application Modeller to ensure they uniquely identify only the single, desired element.
Question 4
A developer needs to expose a Blue Prism process as a secure SOAP web service to be consumed by an external enterprise application. The security requirement is that only authenticated users from the consuming application, which uses basic authentication, can invoke the service. Which sequence of steps correctly configures this secure web service in Blue Prism?
Answer and explanation
Correct answer: B
This is the correct procedure. Security for exposed web services is configured at the Web Service group level in System Manager. You must first create a group, explicitly enable the required authentication mode (in this case, 'Basic Authentication'), and then assign a Blue Prism role that has the 'Web Service Listener' permission. Any processes published under this group will inherit these security settings, ensuring that incoming requests are properly authenticated against Blue Prism's user database.
Question 5
An automation is designed to process high-priority financial transactions that must be completed within a 15-minute SLA. A standard priority queue is used for normal transactions. The solution designer needs to ensure the high-priority items are processed immediately, even if thousands of standard items are already in the queue. The current design uses a single process that gets the next item from the queue. What is the most effective design to handle the high-priority items without disrupting the standard workflow?
Answer and explanation
Correct answer: B
This is the most efficient and standard Blue Prism design pattern for prioritization. By using tags (e.g., 'HighPriority'), the 'Get Next Item' action can be configured with a filter. The process logic should be: 1. Try to get an item with the 'HighPriority' tag. 2. If none is returned, try to get the next available item without a filter. This ensures that the same pool of robots can service both priorities, always favoring the high-priority work without needing separate queues or complex resource management.
Question 6
True or False: When a Blue Prism process is exposed as a web service, the 'max-instances' setting in the web service configuration in System Manager determines the maximum number of concurrent requests the service can handle by spinning up that many process instances.
Answer and explanation
Correct answer: B
This statement is false. The 'max-instances' setting is a throttling mechanism that limits the number of concurrent requests that are queued to be processed. A web service call does not spin up a new process instance on its own. The requests are added to an internal queue and are processed by available runtime resources that are running the target process. It is a limit on inbound requests, not a mechanism for creating new robot sessions.
Question 7
A solution is designed with a 'Controller' process that populates a work queue from a master spreadsheet. Multiple 'Worker' processes then work the items from this queue. The design document states that the Controller process must run to completion before any Worker processes begin, to avoid processing an incomplete set of data. Which mechanism should be used to enforce this dependency in the most reliable way?
Answer and explanation
Correct answer: B
The Blue Prism Scheduler is designed for this exact purpose. It allows for the creation of dependencies between schedules. By making the Worker schedule dependent on the Controller schedule, the system ensures the Worker processes will only be launched after the Controller has finished its run successfully. This is the most robust, manageable, and intended method for orchestrating process dependencies.
Question 8
A retail company is automating its order fulfillment process. The process reads order details from an e-commerce platform's REST API, processes them, and updates an internal inventory system. The development team has created a Visual Business Object (VBO) to handle all API interactions. During testing, they find that the API's authentication token expires every 60 minutes, causing the process to fail with '401 Unauthorized' errors.
The current VBO design has separate actions for 'Get Orders', 'Update Order Status', and other business functions. Each of these actions makes a direct call to the API. There is no centralized mechanism for handling authentication.
What is the most appropriate design change to the VBO to manage the API authentication lifecycle efficiently and robustly?
Answer and explanation
Correct answer: C
This is the correct best practice for managing state, like an authentication token, within a VBO. By using static data items (which retain their value between action calls) and a centralized private action, the VBO can manage its own state. The private 'Get_Auth_Token' action acts as a gatekeeper. It checks if a valid, non-expired token already exists. If it does, it returns it immediately. If not, it performs the authentication, stores the new token and its expiry time, and then returns it. This is efficient (avoids unnecessary logins) and robust (centralizes the logic).
Question 9
A process has been designed to read data from an Excel spreadsheet, but it is failing intermittently with a 'System.OutOfMemoryException' when processing very large files. The current design uses the 'Get Worksheet as Collection' action from the MS Excel VBO to load the entire sheet into memory at the start of the process. The solution must be redesigned to handle files of any size without running out of memory. Which design change is the most appropriate solution?
Answer and explanation
Correct answer: B
This design pattern avoids loading the entire file into memory. By iterating through the spreadsheet row-by-row (or in small batches), the memory footprint of the process remains small and constant, regardless of the file size. While this approach can be slower due to repeated COM interop calls to Excel, it is the only robust way to guarantee that the process will not fail due to memory limitations when handling arbitrarily large files.
Question 10
Case Study: Invoice Processing Automation
Company Background: Global Freight Inc. (GFI) is a logistics company that processes over 5,000 invoices daily. These invoices arrive as PDF attachments in a dedicated email inbox. The finance team manually opens each email, downloads the PDF, extracts key information (Invoice #, Vendor Name, Amount, Due Date), and enters it into their legacy ERP system, which is a Windows-based desktop application.
Automation Requirements: GFI wants to automate this process using Blue Prism. The solution must be scalable to handle peak volumes of up to 10,000 invoices per day. The solution needs to be resilient to application outages and handle various invoice formats from different vendors. A key requirement is that the solution must provide a clear audit trail and a mechanism for the finance team to handle invoices that the robot cannot process (exceptions).
Proposed Solution Design: The proposed solution consists of three processes:
Process 1 (Email Reader): Runs every 15 minutes. It reads the inbox, downloads PDF attachments, and adds a new item to a 'New Invoices' work queue for each PDF. The email is then moved to a 'Processed' folder.
Process 2 (Data Extraction): Multiple robots run this process concurrently. They get an item from the 'New Invoices' queue, use an OCR tool to extract data from the PDF, and if successful, add the extracted data to a 'Validated Invoices' work queue. If OCR fails or confidence is low, the item is marked as an exception in the 'New Invoices' queue.
Process 3 (ERP Entry): Multiple robots run this process. They get an item from the 'Validated Invoices' queue and enter the data into the ERP system. If the ERP system is down, the process should stop gracefully and resume later.
Which of the following represents a critical flaw in this proposed solution design?
Answer and explanation
Correct answer: C
This is a critical flaw. The process creates a new work queue item in 'Validated Invoices' without retaining a link to the original item in 'New Invoices'. If Process 3 fails to enter an item into the ERP, there is no straightforward way to trace it back to the original PDF or email. A robust design would use a single work queue and update the status of the item as it moves through the stages (e.g., 'New', 'OCR Complete', 'ERP Entry Successful'). Alternatively, if multiple queues are used, the item key or ID from the first queue must be passed to and stored in the second queue item to maintain the audit trail.
