Skip to content
This is NOT an official site of the Government of Canada. Click here for the official AI registry.

Helpful Places

Government of Canada – DTPR AI Register

Back to your search

AI Assistant for Government Pay Case Backlog Reduction

Employment & Work · Planning & Decision-making

What it collects that can identify you

Sensitive personal information
Pseudonymous data
  • Synthetic account data modelled on case management tool (CMT) and Phoenix pay system records — including case details such as whether a public servant is being promoted into a higher-paid position. Currently only synthetic data is used; no live personal data is ingested at this stage. Classified up to Protected B.

AIA — Section 3.1, Q51

Also collects operational data, which is anonymized data.

Run by
Public Services and Procurement Canada (PSPC)
Where
No fixed location
Kept
Retained Not specified in the AIA; subject to Government of Canada records management and pay administration retention schedules
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

Employment & WorkPlanning & Decision-making

This AI assistant supports compensation agents at Public Services and Procurement Canada in resolving a backlog of government employee pay cases. It analyzes case data and generates recommendations and summaries to help agents make faster, better-informed decisions — but all final decisions remain with the human agent. Currently, the system operates only on synthetic (non-real) data for acting pay cases older than 365 days.

What it collects and what happens to it

Data taken in

Sensitive personal information
Pseudonymous data
  • Synthetic account data modelled on case management tool (CMT) and Phoenix pay system records — including case details such as whether a public servant is being promoted into a higher-paid position. Currently only synthetic data is used; no live personal data is ingested at this stage. Classified up to Protected B.

AIA — Section 3.1, Q51

Operational data
Anonymized data
  • Approved standard operating procedures (SoPs), job aids, directives, and policy documents that serve as the authoritative knowledge base against which case data is validated. Also includes case selection criteria (case type: acting; age: over 365 days).

AIA — Section 3.1, Q6 and Q23

Processing

Search & Retrieval
  • The system retrieves relevant passages from approved SoPs, job aids, directives, and policy documents in response to case-specific queries, using this retrieval to ground its recommendations in authoritative sources.

AIA — Section 3.1, Q24

Classification & Prediction
  • The system compares account data against knowledge sets to identify patterns, anomalies, and mismatches — classifying case details (e.g. salary entitlements, union dues applicability) to generate structured recommendations for compensation agents.

AIA — Section 3.1, Q23

What it does

Understanding (Semantic AI)
Human decides
  • The system understands and retrieves relevant information from approved standard operating procedures, job aids, directives, and policy documents, matching case data against these knowledge sets to surface relevant guidance for compensation agents.

AIA — Section 3.1, Q21 (Role of the system)

Deciding (Analytical AI)
Human decides
  • The system analyzes account and case data — checking for accuracy, completeness, and compliance with approved SoPs — and scores or classifies cases to generate recommendations. Compensation agents retain full decision-making authority; no decision is automated.

AIA — Section 3.1, Q20–Q21

Outputs

A recommendation or prediction
Anonymized data
  • A detailed case summary and set of recommendations identifying relevant salary information, applicable policy references, and suggested resolution steps for each acting pay case. The output also includes a transparent explanation of the steps taken to arrive at each recommendation, enabling CAs to review, check for errors, and make amendments.

AIA — Section 3.1, Q24

Run by

Public Services and Procurement Canada (PSPC)
  • PSPC Human Capital Management aims to address and eliminate the existing backlog of pay cases by implementing an AI assistant designed to support compensation agents. The department is the accountable deployer of this AI system under the Treasury Board Directive on Automated Decision-Making.

GC HR and Pay - AIA (Open Canada Registry)

Built by

Not stated by the Helpful Places.

Kept for

Retained Not specified in the AIA; subject to Government of Canada records management and pay administration retention schedules
  • An audit trail is maintained recording all recommendations, decision points, overrides, system version used, and change control records. Specific retention duration is not stated in the AIA.
  • Duration: Not specified in the AIA; subject to Government of Canada records management and pay administration retention schedules

AIA — Section 3.2, Q17–Q31

Shared with

Available to the accountable organization
  • Output data (case summaries and recommendations) is available to PSPC Human Capital Management and the compensation agents who use the system. An audit trail records all recommendations, decision points, overrides, and system version used for each decision, accessible to authorized users.

AIA — Section 3.2, Q17–Q25

Not available to me
  • Individual public servants whose pay cases are processed cannot directly access the AI system's outputs or audit trail. The system is an internal tool for compensation agents only; affected individuals interact with outcomes through the normal pay administration process.

AIA — Section 3.1, Q2 (two clients identified: CAs and public servants)

Stored

Stored on 3rd Party Cloud
  • The system interfaces with other IT systems including internet- or telephony-connected devices, suggesting cloud or networked storage. The AIA notes Shared Services Canada involvement, implying Government of Canada cloud infrastructure. Exact storage location and jurisdiction are not specified.
  • Duration: Not specified in the AIA

AIA — Section 3.1, Q47–Q48; Section 3.2 (Shared Services Canada)

How to read the colours

Can it identify you?

Anonymized data
Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data
Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data
The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides
This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes
This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous
This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the Helpful Places, not the vendor.

Your rights

  • Right to a Human ReviewAll recommendations produced by the AI are reviewed by a compensation agent before any action is taken. No decision is automated — CAs retain full decision-making authority and can review the AI's analysis, check for errors, and make amendments as required. The system enables human override of all system outputs.
  • Right to Algorithmic TransparencyFor each recommendation, the AI identifies the steps taken to arrive at it, including which SoPs, job aids, directives, and policies were consulted. The AIA is publicly available, and a general description of the system's logic must be discoverable via a departmental website under the Directive on Automated Decision-Making. The algorithm is not a trade secret.
  • Right to Be Informed of AI UseA plain language notice will be posted through all service delivery channels in use (internet, in person, mail, or telephone) informing clients and affected public servants that an AI system is in operation and contributing to pay case processing. This is required under Impact Level 2 of the Directive on Automated Decision-Making.
  • Right to an Explanation of a DecisionWhere a decision results in the denial of a benefit or service, a meaningful plain-language explanation must be provided, covering: the role of the AI system, the training and client data used and how it was collected, the criteria applied, the output produced, and a justification of the final decision including principal factors. Relevant recourse options must also be communicated. Required under Impact Level 2 of the Directive on Automated Decision-Making.
  • Right to Non-discriminationA Gender-Based Analysis Plus (GBA+) will be conducted and made publicly available to identify and address any discriminatory impacts on gender and other identity factors. Processes will be in place to test datasets against biases and unexpected outcomes. Public servants whose pay cases are processed have the right to equitable treatment regardless of gender, age, disability, or other protected characteristics.

Risks and safeguards

  • Civil liberties harmPotential for discriminatory or biased recommendations if the AI system inherits biases from training data or SoPs, affecting protected groups among public servants with pay cases. The AIA acknowledges the project is in an area of public scrutiny and that clients are particularly vulnerable. Mitigations: A Gender-Based Analysis Plus (GBA+) will be conducted and made publicly available; documented processes will test datasets against biases and unexpected outcomes; internal consultations include the ATIP office and PSPC Privacy; a Privacy Impact Assessment is planned. Currently, only synthetic data is used, limiting live exposure.
  • Psychological harmModerate impact on the health and well-being of compensation agents if automation is perceived as threatening job roles, or if errors in AI recommendations increase cognitive burden and stress. Conversely, there is a positive risk that automation reduces repetitive manual tasks and improves job satisfaction. Mitigations: The hybrid-by-design approach preserves CA decision-making authority; user feedback mechanisms will be in place; the system enables human override; consultation with internal stakeholders was conducted before design.