# crmarena / crmarena_0406_case_routing - taskset: [crmarena](https://harnessreport.com/tasks/crmarena.md) - difficulty: medium - category: crm - language: - runnable from the site: no - agent timeout: 600s ## Results by harness _none yet_ ## Instruction ``` # CRM Assistant Task You are an expert Salesforce CRM assistant helping an internal user (i.e., an employee of the same company). Your job is to answer the question below by reasoning step by step and querying a live Salesforce org using the `simple_salesforce` Python library. ## Approach Think step by step: 1. Identify which Salesforce objects and fields are relevant to the question 2. Write and execute SOQL/SOSL queries to retrieve the data you need 3. Inspect the results, refine your queries if needed 4. Write the final answer to `/app/answer.txt` ## Salesforce Access Connect using the credentials provided as environment variables: ```python import os from simple_salesforce import Salesforce sf = Salesforce( username=os.environ["SF_USERNAME"], password=os.environ["SF_PASSWORD"], security_token=os.environ.get("SF_SECURITY_TOKEN", ""), domain=os.environ.get("SF_DOMAIN", "login"), ) ``` Use SOQL for structured queries and SOSL for full-text search: ```python # SOQL — structured query result = sf.query_all("SELECT Id, Subject FROM Case WHERE Status = 'Open'") records = result["records"] # SOSL — full-text search across objects result = sf.search("FIND {Smith} IN NAME FIELDS RETURNING Contact(Id, FirstName, LastName)") records = result["searchRecords"] ``` ## Available Salesforce Objects The org contains these 16 objects (key fields listed): | Object | Key fields | |---|---| | `User` | Id, Alias, FirstName, LastName, Email, Phone | | `Account` | Id, FirstName, LastName, PersonEmail, Phone, ShippingCity, ShippingState, RecordTypeId | | `Contact` | Id, FirstName, LastName, Email, AccountId, OwnerId | | `Case` | Id, Subject, Description, Status, Priority, ContactId, AccountId, CreatedDate, ClosedDate, OwnerId, IssueId__c, OrderItemId__c | | `Knowledge__kav` | Id, Title, Summary, FAQ_Answer__c, UrlName | | `ProductCategory` | Id, Name, CatalogId | | `Product2` | Id, Name, Description, IsActive, External_ID__c | | `ProductCategoryProduct` | Id, ProductCategoryId, ProductId | | `Pricebook2` | Id, Name, Description, IsActive, ValidFrom, ValidTo | | `PricebookEntry` | Id, Pricebook2Id, Product2Id, UnitPrice | | `Order` | Id, AccountId, Pricebook2Id, Status, EffectiveDate | | `OrderItem` | Id, OrderId, Product2Id, PricebookEntryId, Quantity, UnitPrice | | `EmailMessage` | Id, Subject, TextBody, FromAddress, MessageDate, ParentId, ToIds | | `LiveChatTranscript` | Id, Body, AccountId, ContactId, CaseId, OwnerId, EndTime, LiveChatVisitorId | | `Issue__c` | Id, Name | | `CaseHistory__c` | Id, CaseId__c, Field__c (one of: 'Case Creation', 'Owner Assignment', 'Case Closed'), NewValue__c, OldValue__c, CreatedDate | ## Task-Specific Instructions ## Case Routing Policy The case routing policy determines the best agent to assign the given new case based on the following criteria - Issue Expertise: The agent who has closed the most cases associated with the issue most similar to the given case. - Product Expertise: If there is a tie in issue expertise, the best agent is the one who has solved the most cases associated with the product most relevant to the given case. - Workload: If there is still a tie, the best agent is the one that has least cases with Status not 'Closed'. ## Additional Context # Domain Details ## Quarters of the Year - Q1: January 1 to March 31 (both inclusive). - Q2: April 1 to June 30 (both inclusive). - Q3: July 1 to September 30 (both inclusive). - Q4: October 1 to December 31 (both inclusive). ## Seasons - Winter: December 1 to February 28/29 (both inclusive). - Spring: March 1 to May 31 (both inclusive). - Summer: June 1 to August 31 (both inclusive). - Autumn/Fall: September 1 to November 30 (both inclusive). ## Time Periods - Past 2 quarters: This refers to any timeframe spanning two quarters back from a specified `end_date`. That translates to a six-month period retrospectively from the `end_date`. - Issue Significantly More Than Other Months: This means there is a month where the number of cases reported are larger than all other months. ## Question Use the case routing policy to determine the most suitable agent for the given case. Return only the Id of the agent. ## Given Case - Case Subject: Size Issue with HydroShield Swim Cap - Case Description: I purchased the HydroShield Swim Cap based on the size chart provided on your website, but the actual product size does not match the chart. The cap is either too tight or too loose, making it uncomfortable to use. Can you please help me resolve this size mismatch issue? ## Output Requirements Write your final answer to `/app/answer.txt`: - Single value: write it directly (e.g., `Agent A`) - Multiple values: comma-separated (e.g., `Agent A, Agent B, Agent C`) - No results: write `None` - Answer only — no explanations, no sentences, no punctuation beyond commas ```python from pathlib import Path Path("/app/answer.txt").write_text("your answer here") ``` ``` --- Harness Report runs agent harnesses from their GitHub repos on Harbor tasks and records every model call. Every page is also `.md` and `.json`; index: https://harnessreport.com/llms.txt · MCP: https://harnessreport.com/mcp