Architecture & capabilities

Overview

The Job Rate Manager is built on four agentic skills that together take a work order or project transaction from intake through to an updated IFS Cloud record: Work Transaction Intake, Job Classification & Rule Evaluation, Time Categorization, and Exception Handling. This article covers those skills, the workflow they share, integration points, and how the system keeps a person in the loop when information is missing or rules conflict.

Agentic Skills

Work Transaction Intake

Captures work order and project transaction details, including employee information, job type, location, and the labor information required to determine a rate. This skill extracts the transaction details that every downstream skill depends on, and is what triggers the Job Rate Manager the moment a new transaction comes in.

Job Classification & Rule Evaluation

Identifies the applicable union agreements, equipment, shift, and labor classifications for the transaction, and validates the labor conditions against them. This skill matches the applicable customer, location, and job rules, and loads the geography-specific rules, calendars, and rate agreements that govern the entry.

Time Categorization

Applies business rules, premiums, overtime, allowances, and customer-specific conditions to categorize time into the correct time type. This skill splits reported hours into regular, overtime, and double-time categories by threshold, and recommends the correct time and cost types.

Exception Handling

Identifies missing information, conflicting rules, or ambiguous scenarios that require human review. Nothing is finalized without review when a case is flagged: the worker surfaces what’s missing or conflicting and routes the case to a person rather than resolving it on an assumption.

Job Rate Workflow

The following table outlines the process from transaction intake through to a billing and payroll update.

Stage	What the System Does	Action / Output	Channel
Work Request Intake	Extracts work details, employee, equipment, and job type	Required inputs validated	ERP
Rule Identification	Matches applicable union, customer, location, and job rules	Correct rule set determined	ERP
Labor Time Classification	Applies premiums, overtime, allowances, and classifications	Time types calculated	Rules Engine
Exception Validation	Detects missing or conflicting information	Exceptions escalated	Teams
Rate Confirmation	Validates the calculation	Audit trail generated	ERP
Billing & Payroll Update	Prepares final outputs	Payroll and billing updated	ERP

Integration Points

IFS Cloud: The primary system of record. The Job Rate Manager reads work order and project transaction details and writes back by splitting, creating, or updating time and cost transactions once classification and rule evaluation are complete.

Rules Engine: Applies labor, union, travel, equipment, fringe, and burden cost rules against the classified time, and is the authority classification is drawn from before any update is made.

Collaboration (Microsoft Teams): Used to escalate exceptions, missing information, and conflicting rules to a person for review.

Payroll systems: Job cost details are prepared in a form ready for export to ADP and other payroll systems.

Human-in-the-Loop Design

The Job Rate Manager is designed so that classification and rule application run automatically, while genuine ambiguity is routed to a person rather than resolved on a guess.

The system flags for review when:

• Information required to determine a rate is missing from the transaction

• Applicable rules conflict with one another

• A scenario is ambiguous enough that finalizing it without review carries risk

The system proceeds automatically when:

• The applicable agreement, classification, and rule set can be determined cleanly from the enriched transaction data

• Hours can be classified against regular, overtime, and double-time thresholds without contradiction

In every case, classification comes from the rule engine first, and the update to IFS Cloud follows only once that classification is set. This ordering is what makes the output safe to hand to payroll: nothing is written based on the worker’s own judgment ahead of the rules that govern it.

Feedback and Continuous Improvement

Because rate determination runs on a rule engine rather than case-by-case judgment, the same agreement and the same set of facts produce the same classification every time, regardless of who entered the original data. Every rate decision generates an audit trail as part of rate confirmation, giving payroll and finance teams a consistent, traceable record of how each transaction was classified and why, rather than a rate that depended on which person happened to calculate it.