Skip to main content

What AI Is Your Council Already Using? Start with a Simple Register.

A proportionate AI register helps council teams see which tools and tasks are being tested, who owns them, what information is involved and when review is due—without mistaking an inventory for approval or a full risk assessment.

Author avatar
Steve Agi 13 November 2026 · 7 min read
What AI Is Your Council Already Using? Start with a Simple Register.

A useful question before the next AI trial

A council may approve a new AI product and still have a partial picture of how AI is being used. A writing assistant inside office software, a transcription feature in a meeting platform and a staff member testing a public chatbot may all involve AI, but they can sit in different teams and follow different rules.
Without a shared view, it is difficult to see who owns each use, what information it touches, whether another team is already testing something similar or what needs to be reviewed before the use changes. The National AI Centre recommends that organisations maintain an AI register covering their AI systems, use cases, accountable people and appropriate governance level. It says the register can help reveal existing tools, reduce duplication and support shared understanding. It also warns that a register does not replace impact assessment or risk management. 
This guide suggests a proportionate starting point for council teams. It is not a statutory template. The Australian Government’s mandatory AI policy applies to covered Commonwealth entities, not automatically to local councils; councils should check their own state, territory and organisational requirements. 

Start with the use, not just the product

A list of product names is not enough. A single product may be used for several different tasks, and those tasks can involve different people, information and consequences. Record both the tool and the use case: the task the AI supports, the business purpose and what happens to its output.
Include AI that is being considered, tested, used or retired. Remember features built into software the council already owns. The National AI Centre specifically recommends including embedded AI features.  If the team is unsure whether a feature counts, mark it for review rather than leave it out of sight.
The point is not to catalogue every prompt or collect copies of staff work. It is to provide enough structured information to find the use, understand its purpose, locate its owner and decide what governance applies.

A practical starter register

A council can begin with one controlled spreadsheet or an existing governance register. Use fields that answer these questions:
•What is the use? Give it a short name, a reference number and a plain-language description. State the business objective and what the AI does.
•Which tool is involved? Record the product or embedded feature, vendor if known, and broad AI type. Note whether the use is being explored, tested, operated or retired.
•Who owns it? Name the business owner or responsible role and the team. A technical contact may also be needed, but ownership should sit with someone who understands the work the use supports.
•Who or what could be affected? Note the intended users, the process involved and whether the output may affect residents, employees, suppliers, service access or decisions.
•What kind of information is involved? Use a high-level classification such as public, internal or sensitive, and record the authorised handling rule or policy reference. Do not copy personal or confidential source material into a broadly accessible register.
•What review has happened? Link to the relevant privacy, security, procurement, records, legal or impact assessment in the approved system. Record open questions and the status of approvals. A register entry is not permission to use a tool.
•What human oversight applies? State who checks the output, what they check and who remains accountable for any material that is used or shared.
•How will the council know whether it is useful? Record the expected benefit and one or two measures, such as review effort, correction rate or completion time, with quality safeguards where needed.
•When does it need another look? Capture the status, last-review date and a review trigger, such as a change in purpose, data, user group, product capability or vendor terms.
The Australian Government’s register standard for covered agencies includes fields such as a use-case description and business objective, AI type, lifecycle stage, accountable owner, risk ratings and the date the assessment was last updated. That is a useful design reference, not a council compliance checklist. 

Keep the register proportionate

A register should make oversight easier, not create a parallel approval bureaucracy. Use a council’s existing technology, procurement, privacy and records processes wherever possible. Store the register in a controlled location, give it an owner and avoid duplicating detailed assessments that already live in approved systems.
If a field is unknown, write “to confirm,” name who will confirm it and set a review date. A blank field can look like the question has been answered when it has not. If the council has only a few low-risk uses, a basic inventory may be enough to start. A use that could influence decisions about people, involve sensitive information or affect essential services needs more detailed assessment and oversight under the council’s applicable rules.
The register should not be treated as a scorecard that labels every use “low risk” or “approved.” A risk rating depends on the actual task, information, people affected and operating conditions. It should be assessed by the people and processes authorised to do that work, not inferred from a short spreadsheet entry.

Use the register to answer operational questions

A useful register helps staff answer practical questions before they buy, enable or expand an AI feature:
•Is this tool or a similar use already being tested elsewhere in the council?
•Who can explain why it is being used and what boundaries apply?
•Has the team checked whether the information and product are approved for this purpose?
•What assessment or approval is still outstanding?
•If the output changes a report, recommendation or service, who checks and owns the final result?
•What would cause the council to revisit the use?
The National AI Centre describes the register as a way to help staff and stakeholders understand where and why AI is being used, discover existing systems and reduce duplication. It also says that a register supports conversations about AI; it does not replace deeper governance. 

Example: a small, visible trial

Consider a hypothetical trial in which a governance team uses an approved tool to draft a plain-language summary of a publicly available council report for staff review. The register could identify the task and intended benefit, the tool and trial stage, a responsible business owner, the public-information boundary, the reviewer and the quality checks, the assessment reference, the trial measure and the review date.
It would not need to store the report, prompt or draft summary in the register. Those materials should be handled under the council’s information and records rules. If the team later expands the task to unpublished papers, adds a public-facing chatbot or changes how the summary informs decisions, that is a material change to assess—not a minor edit to the product name.

Keep it current as AI use changes

A register becomes unreliable if it is set up once and forgotten. Ask the owner to update it when the task moves from exploration to trial or live operation, when the product or vendor changes materially, when new information or users are involved, or when the use stops.
For covered Commonwealth agencies, the current policy requires in-scope use cases to be registered, monitored and evaluated, and their assessments revalidated when scope, usage or operation materially changes. These obligations do not automatically apply to councils, but the review triggers illustrate why a register should track more than a tool name. 
A brief periodic review can check whether entries still have an owner, whether status and approvals are current, whether reported benefits are supported and whether any use should be retired or assessed further. The review frequency should reflect the scale and risk of the use and the council’s own governance requirements.

A practical first step

Start by asking a small cross-functional group—business, digital or IT, privacy, records and procurement—to list the AI tools and uses they already know about. Include embedded features and active trials. Do not begin by asking staff to disclose prompt histories or upload case material.
Agree on a simple template, an owner and a place for the register. Test it against a few known uses. Remove fields no one can act on, add references to the council’s existing approval routes, then invite other teams to identify uses the first group may not see. Use the resulting inventory to decide what needs an assessment, a staff conversation, a duplicate purchase check or no further action.
A register will not make an AI use safe by itself. It gives a council a clearer starting point: what is in use, why it is there, who is responsible and what needs to happen next.

References

[1] National AI Centre, “AI systems register.”

[2] Australian Government Digital Transformation Agency, “Standard for accountability.”

Published by

Steve Agi Autonomous Automation Architect, C51 Consulting

About our partner

C51 Consulting

C51 Consulting helps Australian councils and public sector teams improve reporting, strengthen evidence traceability and reduce avoidable administrative effort. Council Reporting & Evidence Assurance — Review a defined corporate reporting process to identify evidence gaps, manual follow-up and unclear ownership. Indicative diagnostic: A$7,500 + GST, subject to agreed scope. Public Sector Workflow Reviews — Map one operational workflow, pinpoint bottlenecks and duplicated work, and identify practical improvements and appropriate automation options. Indicative diagnostic: A$7,500 + GST, subject to agreed scope. Practical AI Productivity Clinics — Build staff capability through a practical session on useful tasks, responsible use and internal policy constraints. Indicative clinic: A$1,500 + GST, subject to format and agreed scope. We start with a defined business problem, agree deliverables and focus on measurable improvement opportunities. We work alongside existing systems where practical; implementation is separately scoped. Clinics respect each organisation’s security, privacy, records-management and AI-use requirements. Request a scoping conversation about a council reporting process, government workflow or team clinic. https://c51consulting.com

Learn more