Service-catalog requests & low-code flows
Fulfil catalog requests and build approval flows in enterprise low-code builders.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- ServiceNow · Salesforce Flow Builder · Power Automate
- Class GUI-need label
G3· GUI-only
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R05.05.01.04-T01 Order three configured developer laptops from the ServiceNow portal
Original instruction
ServiceNow developer instance, Service Portal (/sp) as user 'abel.tuter'. Order 3 x 'Developer Laptop (Mac)' with 32 GB RAM, 1 TB SSD and the 'Adobe Creative Cloud' add-on, cost center CC-410, business justification from input/justification.txt. Submit and record the REQ number.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
sc_req_item: one RITM with quantity 3 and variables equal spec; requested_for abel.tuter; REQ number matches output. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
50% variables correct; 20% quantity; 20% cost center/justification; 10% number
R05.05.01.04-T02 Report who is holding up a pending catalog approval
Original instruction
ServiceNow developer instance, Service Portal as 'abel.tuter'. Request RITM0010042 is still pending. Find its current stage and the name of the approver whose approval is outstanding, and write 'stage;approver' to output/answer.txt.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicknav· Pan/zoom/scroll- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 1 / 5
Proposed verifier
Exact match on stage and approver from sysapproval_approver (one approver already approved, a decoy). REST/SOAP is ACL-denied and the agent lacks the itil role, so the approval is reachable only via the portal.
Proposed partial credit
50% stage; 50% approver
R05.05.01.04-T03 Raise a normal change request with risk, CI and maintenance window
Original instruction
ServiceNow developer instance, ITIL user. Create a Normal change: short description 'Patch mail cluster to 3.4.2', CI 'mail-prod-01', risk Moderate, impact 2, planned window 2026-10-24 01:00-03:00 (instance timezone US/Pacific), implementation and backout plans from input/change_plan.docx. Submit for assessment.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
change_request: type normal, cmdb_ci, risk, impact, start/end dates in UTC equal key, plans populated, state Assess. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
40% classification fields; 30% window with correct timezone; 20% plans; 10% state
R05.05.01.04-T04 Submit a new-hire order guide for a finance analyst
Original instruction
ServiceNow developer instance. Using order guide 'New Hire Setup', onboard 'Priya Nair' (Finance, start 2026-11-02, manager 'Beth Anglin'): standard Windows laptop, 27-inch monitor, Office license, access to group 'FIN-AP' and shared drive '\\fin\ap'. Submit and check that all generated RITMs exist.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
sc_request with RITMs per key item set; each RITM's variables (start date, manager, group) correct; no extra items. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
40% correct RITM set; 40% variables; 20% no extra items
R05.05.01.04-T05 Submit a password-reset request on a user's behalf via impersonation
Original instruction
ServiceNow developer instance, admin. Impersonate 'beth.anglin', open catalog item 'Password Reset' in the portal, choose system 'SAP ERP', verification method 'Manager approval', and submit. End impersonation, then add a work note on the resulting RITM: 'Submitted on behalf per call INC0012011'.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
RITM opened_by and requested_for = beth.anglin via impersonation (sys_audit impersonation records); variables correct; work note by admin after impersonation ended. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
40% request under impersonation; 30% variables; 30% work note by admin
R05.05.01.04-T06 Build and activate a record-triggered case escalation flow in Flow Builder
Original instruction
Salesforce developer org. In Flow Builder, create a record-triggered flow on Case (after save, created or updated) that runs when Priority changes to High: set Escalated__c = true, assign to queue 'Tier 2', and post a Chatter message to group 'Support Escalations' with the case number. Activate it.
- Platforms
web· Any OS (web)- Applications
- Salesforce Flow Builder
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Grader creates 3 test cases (High on create, change to High, Medium) and checks field values, owner and Chatter posts; active flow version exists. API Enabled is removed from the agent profile, so Metadata/Tooling API deploys are denied; Setup Audit Trail must show the change via Flow Builder/Setup UI. Control pair: T14.
Proposed partial credit
20% trigger and entry condition; 40% test outcomes; 20% Chatter post; 20% activated
R05.05.01.04-T07 Route vendor PDF attachments to SharePoint with a Power Automate cloud flow
Original instruction
Power Automate (maker portal, test tenant). Build a cloud flow 'AP intake': when an email arrives in ap@contoso.test from any @vendor.test address with attachments, save only PDF attachments to SharePoint site 'Finance' library 'AP/inbox' named '<yyyyMMdd>_<sender>_<filename>', then mark the email read. Turn it on.
- Platforms
web· Any OS (web)- Applications
- Power Automate · SharePoint · Outlook
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Grader sends 4 test emails (PDF from vendor, PNG from vendor, PDF from other domain, 2 PDFs from vendor): SharePoint contains exactly the key files with correct names; emails marked read accordingly. No app registration or pac CLI auth is issued, Conditional Access admits the agent only via the sandbox browser, and solution import is denied; flow must be authored in the maker portal (designer edit history).
Proposed partial credit
20% trigger filter; 50% test outcomes; 20% file naming; 10% flow on
R05.05.01.04-T08 Create a request type with a form and approval step in Jira Service Management
Original instruction
Jira Service Management Cloud test site, project 'IT Help'. Create request type 'Software purchase' with fields Software name, Cost (number), Cost center (select from input/cost_centers.txt) and Business reason; add workflow approval by the reporter's manager when Cost > 500 using an automation rule. Publish in the portal group 'Purchasing'.
- Platforms
web· Any OS (web)- Applications
- Jira Service Management
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsdrag· Drag & drop- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Request type and form fields exist; grader submits two test requests (300 and 900) and checks approval created only for 900 with the right approver; portal group set. No API token issued.
Proposed partial credit
30% request type and form; 40% approval behavior on tests; 20% automation rule; 10% portal group
R05.05.01.04-T09 Add a finance-review step to an opportunity approval process
Original instruction
Salesforce developer org. Edit approval process 'Approve Opportunity Amount' to add step 2 'Finance review' that applies when Amount > 100,000, approver queue 'Finance Approvers', with rejection returning to the previous step and a final-approval field update Approved__c = true. Keep the process active.
- Platforms
web· Any OS (web)- Applications
- Salesforce Setup
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Grader submits two test opportunities (80k, 150k) and approves/rejects via admin session; ProcessInstanceSteps follow key path; field updates applied. API Enabled is removed from the agent profile, so Metadata/Tooling API deploys are denied; Setup Audit Trail must show the change via Flow Builder/Setup UI.
Proposed partial credit
30% step criteria; 20% approver; 30% test path behavior; 20% final actions
R05.05.01.04-T10 Build a filtered Zap from form responses to CRM and chat
Original instruction
Zapier test account with connected Typeform, HubSpot (test portal) and Slack (test workspace). Create a Zap: new Typeform response on form 'Demo request' -> filter company size >= 50 -> create/update HubSpot contact (email, company, size) -> post to #sales-leads with name and company. Turn it on and run the 3 test submissions in input/test_responses.json.
- Platforms
web· Any OS (web)- Applications
- Zapier · HubSpot · Slack
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP06· Code/engineering- Conditions
multiapp· Multi-app- Proposed difficulty
- 3 / 5
Proposed verifier
HubSpot contains exactly the 2 qualifying contacts with fields; Slack channel has exactly 2 messages with correct text; Zap is on. Zapier, HubSpot and Slack API credentials are withheld from the agent; contacts and messages must be attributed to the Zapier app (HubSpot source, Slack bot_id) and match Zap task-history runs.
Proposed partial credit
20% filter; 40% HubSpot records; 30% Slack posts; 10% Zap on
R05.05.01.04-T11 Create a catalog item with variables and a UI policy in ServiceNow
Original instruction
ServiceNow developer instance, admin. Using Catalog Builder or the classic form, create item 'Conference Room Equipment' in category Hardware: variables Room (reference to cmn_location), Equipment (multiple choice: Projector/Display/Speakerphone), Delivery date, Notes; a UI policy hiding Notes unless Equipment = Display; publish to the Service Portal.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
sc_cat_item and item_option_new rows equal spec; catalog UI policy with condition and action; grader opens the portal as a user and verifies Notes visibility toggles. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
30% item and category; 40% variables; 30% UI policy behavior
R05.05.01.04-T12 Fulfil six catalog tasks and close their requests within SLA
Original instruction
ServiceNow developer instance as fulfiller in group 'Hardware'. Work the 6 open SCTASKs assigned to the group: for laptops, record asset tag and serial from input/asset_list.xlsx in the closure variables and assign the asset to the requester; for monitors, mark delivered. Close tasks so all RITMs reach Closed Complete before SLA breach.
- Platforms
web· Any OS (web)- Applications
- ServiceNow · Excel
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP08· App/system state- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
sc_task closed with variables equal key; alm_hardware assets assigned to correct users; RITMs closed complete; task_sla records not breached. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
40% closure variables; 30% asset assignments; 20% RITMs closed; 10% SLA
R05.05.01.04-T13 Record a desktop RPA flow that keys invoices into a legacy app
Original instruction
Power Automate Desktop (Windows). Build flow 'KeyInvoices' that reads input/invoices.xlsx and, for each row, fills the 'LegacyAP' WinForms app (Vendor, Invoice #, Date, Amount) using UI elements, clicks Save and logs the confirmation number back to column F. Run it once over the 10 rows.
- Platforms
win· Windows- Applications
- Power Automate Desktop · Excel
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G3· GUI-only- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
multiapp· Multi-app- Proposed difficulty
- 4 / 5
Proposed verifier
LegacyAP DB has 10 invoices equal the workbook; column F holds the 10 confirmation numbers; the saved .pad flow contains UI-element actions (not SendKeys-only); flow run history shows one successful run. LegacyAP's database is firewalled from the agent and each row's entry time must fall inside the logged PAD run, so direct inserts or hand keying are detected.
Proposed partial credit
40% invoices entered; 25% confirmations logged; 25% UI-element selectors used; 10% run history
R05.05.01.04-T14 Deploy the same case-escalation flow as metadata (control pair of T06)
Original instruction
Control pair of T06. Separate developer org with an API-enabled user (sf CLI authenticated). Write the Flow metadata XML for the same record-triggered escalation logic and deploy it active with 'sf project deploy start'.
- Platforms
lin· Linuxmac· macOSwin· Windows- Applications
- Salesforce CLI · VS Code
- Interfaces
term· Terminalapi· API/SDK- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
AP· API-sufficient- Interaction primitives
code· Code & commands- Work product
WP06· Code/engineering- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Same 3 behavioral test cases as T06; scores compared with T06 to measure the GUI penalty.
Proposed partial credit
20% deploys active; 60% test outcomes; 20% Chatter post
R05.05.01.04-T15 Build a two-level software approval flow in ServiceNow Flow Designer
Original instruction
ServiceNow developer instance. In Flow Designer, create flow 'Software Request Approval' triggered by catalog item 'Software Request': ask the requester's manager for approval; if approved and Cost > 500, ask group 'Finance Approvers'; on rejection at any stage set RITM to Closed Incomplete and notify the requester; on full approval create an SCTASK for 'Software Asset Mgmt'. Activate and attach it to the item.
- Platforms
web· Any OS (web)- Applications
- ServiceNow Flow Designer
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Grader submits 3 test requests (cost 200 approve; 900 approve/approve; 900 approve/reject) and checks approvals, RITM states, SCTASK creation and notifications. sys_hub_flow authored in designer by agent. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
20% trigger/binding; 50% three test paths; 20% notifications; 10% active
R05.05.01.04-T16 Build a three-screen Power Apps canvas app over a SharePoint list
Original instruction
Power Apps (maker portal, test tenant). Create canvas app 'Equipment Loans' on SharePoint list 'Loans': browse screen with search and filter by status, detail screen, and edit form with validation (Return date >= Loan date, required borrower). Add a button that sets Status = Returned and stamps today's date. Publish and share with group 'Lab Staff'.
- Platforms
web· Any OS (web)- Applications
- Power Apps · SharePoint
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & formsgeom· Precise geometry- Work product
WP06· Code/engineering- Conditions
licensed· Licensed pro app- Proposed difficulty
- 4 / 5
Proposed verifier
Grader drives the published app with Playwright: search/filter results, validation rejects bad dates, Return button updates the list item; app shared to group. No app registration or pac CLI auth is issued, Conditional Access admits the agent only via the sandbox browser, and solution import is denied; the app must be authored in Power Apps Studio (app version history).
Proposed partial credit
25% browse/search; 30% form validation; 30% return button; 15% published and shared
R05.05.01.04-T17 Build a Make scenario with a router, iterator and error handler
Original instruction
Make (test account). Scenario 'Order sync': webhook receives an order JSON; iterate line items; route lines with SKU prefix 'DIG-' to a Google Sheet 'Digital' and others to Airtable 'Physical'; on Airtable error, retry 3 times then send an email to ops@example.test. Turn on and replay the 3 payloads in input/payloads/.
- Platforms
web· Any OS (web)- Applications
- Make · Google Sheets · Airtable
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
multiapp· Multi-app- Proposed difficulty
- 4 / 5
Proposed verifier
Sheet and Airtable contain exactly the key rows after replay; the payload with a malformed field triggers the error route (one email in test inbox); scenario blueprint shows router, iterator, error handler modules. Make API tokens and Google/Airtable credentials are withheld from the agent, so rows must be written by the scenario's replay runs (Make execution log).
Proposed partial credit
20% iterator; 35% routing results; 30% error handling; 15% scenario active
R05.05.01.04-T18 Run an offboarding through an order guide and its fulfilment tasks
Original instruction
ServiceNow developer instance. Offboard user 'jdoe' (last day 2026-10-31) via order guide 'Employee Offboarding': disable the account in the sandbox directory catalog task, remove him from the 6 groups listed, create an asset-return task for his 2 assets, reassign his 9 open incidents to his manager, and close all fulfilment tasks.
- Platforms
web· Any OS (web)- Applications
- ServiceNow
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
risk· Irreversible/riskylong· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
sys_user jdoe inactive and locked out; sys_user_grmember rows removed for exactly 6 groups; return task exists referencing 2 assets; 9 incidents reassigned; all SCTASKs closed. REST/SOAP and import sets are ACL-denied for the agent; records must come from interactive portal/form transactions (syslog_transaction).
Proposed partial credit
20% account disabled; 25% group removals; 20% asset return; 35% incident reassignment and tasks closed
R05.05.01.04-T19 Build a parallel approval flow with escalation and Teams adaptive cards
Original instruction
Power Automate (test tenant). Create flow 'Capex approval' on new SharePoint 'Capex' list items: parallel approvals from Finance and Operations via Teams adaptive cards; if either has not responded in 2 days (use a configurable timeout, set 5 minutes for testing) escalate to their manager; write outcome and approvers back to the item. Run 3 grader-scripted test items.
- Platforms
web· Any OS (web)- Applications
- Power Automate · Microsoft Teams · SharePoint
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
licensed· Licensed pro applong· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Test harness responds to cards per script: outcomes on the 3 items equal key (approved, rejected, escalated-then-approved); escalation fired only on timeout; flow run history consistent. No app registration or pac CLI auth is issued, Conditional Access admits the agent only via the sandbox browser, and solution import is denied; flow must be authored in the maker portal (designer edit history).
Proposed partial credit
25% parallel branches; 30% escalation logic; 30% outcomes written back; 15% cards in Teams
R05.05.01.04-T20 Build a screen flow with a loop, subflow and fault paths in Flow Builder
Original instruction
Salesforce developer org. Create screen flow 'Bulk Renewal Quote' launched from an Account: list active contracts expiring in 90 days, let the user pick contracts, loop to create one renewal Opportunity per contract via an autolaunched subflow (uplift 5%), show a summary, and route any DML fault to a screen and a Platform Event 'Flow_Error__e'. Add it to the Account page.
- Platforms
web· Any OS (web)- Applications
- Salesforce Flow Builder
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickdrag· Drag & droptype· Text & forms- Work product
WP06· Code/engineering- Conditions
licensed· Licensed pro applong· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Grader runs the flow via UI test harness on 2 accounts (one with a validation-rule failure): opportunities created with amounts x1.05, fault path publishes the event; Lightning page includes the flow. API Enabled is removed from the agent profile, so Metadata/Tooling API deploys are denied; Setup Audit Trail must show the change via Flow Builder/Setup UI.
Proposed partial credit
20% selection screen; 35% loop/subflow results; 25% fault handling; 20% page placement
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.