Opening Remarks and Audience AI Maturity Polling
Jennifer Arnold, Public Sector Network
- Queensland agencies are moving from experimentation towards contained production use cases, but maturity remains uneven. Leaders need a clear view of organisational readiness before expanding AI into higher-impact services.
- Productivity remains a common starting point, while service quality, workforce support, trust and assurance are becoming equally important measures of value.
- Training, approved tools, practical use cases and access to expertise are necessary if staff are to move beyond isolated pilots and use AI responsibly in daily work.
Ministerial Address
Honourable Steven Minnikin MP, Queensland Government
- Queensland's digital agenda is grounded in service choice. Online channels, digital identity and AI should improve access without removing face-to-face support for people who need it.
- AI should strengthen the work of public servants rather than replace them, with repetitive tasks reduced so staff can focus on service, judgement and human interaction.
- Brisbane 2032 provides a fixed deadline for joined-up digital services, stronger data sharing and more consistent customer experiences across government.
A Vision for Digital Government in Queensland
Mark Howard, Department of Customer Services, Open Data and Small and Family Business
The Hon. Victor Dominello, Future Government Institute (Moderator)
- Queensland is organising digital government around government-to-community, government-to-business and government-to-government value streams, with a horizontal layer supporting rather than replacing agency ownership.
- Shared data, legacy modernisation and responsible AI require a collective digital voice across government, particularly when priorities, funding and capability extend beyond one department.
- Digital investment must work at two speeds: rapid delivery of current government priorities and sustained reform of platforms, assurance, capability and operating models.
A Blueprint for Public Sector AI Adoption
Joel Macfarlane, Datacom
- Moving beyond copilots and grassroots experiments requires foundations across governance, risk, workforce learning, safe tools, modern data and agent management.
- Agencies should prepare for AI to reshape citizen channels, software delivery and back-office work, while keeping human permission and accountability for consequential actions.
- AI adoption introduces new cost pressures as well as productivity gains. Leaders need visibility of agent activity, permissions and consumption before use expands across the organisation.
Beyond Silos: Enabling Cross-Agency Collaboration for Stronger Public Value
Joanne Greenfield, Australian Digital Health Agency
Luke Twyford, Queensland Family and Child Commission
Melissa Wilson, Department of Primary Industries, Queensland Government
Pia Andrews, Northern Territory Government
- Cross-agency collaboration works best when agencies organise around shared life journeys and public outcomes, rather than asking people to navigate institutional boundaries.
- Secure system-to-system information exchange can reduce fragmented case management in health, child safety and biosecurity, but safeguards must reflect the purpose and potential harm involved.
- Multidisciplinary teams, delegated authority and collaboration on unfinished work are practical conditions for joined-up delivery. Technology alone cannot compensate for fragmented ownership or leadership.
Moving from AI Proofs of Concept to Trusted Scale
Gisele Kapterian, Salesforce
- Many AI pilots stall because organisations over-engineer experimentation or underestimate the governance, permissions, observability and escalation needed for enterprise use.
- Assurance must continue through the full workflow between staff, AI systems and community members. Initial approval is not enough when models, data and operating conditions can change.
- Shared responsibility should be explicit across government, platform providers, delivery partners and users, with controls matched to the type of AI capability being procured.
Welcome from Track Chair: Digital Services, Customer Experience and AI Innovation
John Mackenney, Adobe
- Government now needs to design for people and AI agents. Content must remain authoritative, current, structured and accessible if services are to be found and trusted through AI-mediated channels.
- Scaling AI depends on workflow redesign, accountable ownership and measurement, not a larger collection of pilots. Leaders need to define the service or public outcome before selecting the technology.
- Moving from gates to guardrails can support faster delivery, provided teams retain clear principles, evidence, workforce capability and continuing accountability.
Building Organisational Conditions for Innovation and AI Adoption
Adrian Clutterbuck, Children's Health Queensland Hospital and Health Service
- Frontline teams often need permission, mentoring, pathways to expertise and practical delivery support more than they need large innovation budgets.
- Safe experimentation requires freedom within a framework: clear guardrails, a suitable risk appetite and leaders who can hold uncertainty while evidence develops.
- Successful pilots need an agreed route to enterprise funding, procurement, governance and benefits measurement. Without that pathway, useful frontline ideas can remain trapped in the experimental stage.
From Human in the Loop to Human in Control
Dhiraj Chaudhry, Infosys
Paul Beardsell, Infosys and WongDoody
- A nominal human-in-the-loop model can still fail when the person only checks whether AI followed its programmed steps rather than assessing the complete situation.
- AI should gather evidence at scale and disclose missing, uncertain or outdated information, while people retain judgement, empathy and responsibility for the decision.
- Agencies should design each workflow around three questions: what AI should do, what the human should do, and how the process will keep the human genuinely in control.
Queensland 2032 as a Catalyst: Reimagining Public Sector CX and Digital Government for a Global Stage
Rob Bollard PSM, Pega
Lee Stevens, Brisbane 2032
Lea Diffey, former Queensland Government executive
John Mackenney, Adobe, moderator
- Joined-up services need orchestration across existing agency systems, other levels of government and third parties, rather than assuming wholesale replacement of the current technology estate.
- Brisbane 2032 creates a deadline for reliable, secure and contextual services, but agencies should begin with current problems and deliver useful value before the event.
- AI use in regulated services must be predictable, auditable and cost-aware. Constrained workflows with human review are more suitable than uncontrolled networks of autonomous agents.
Transforming Rural Healthcare: Integration, Innovation and Frontline Empowerment
Dr Marlow Coates, Torres and Cape Hospital and Health Service
- In remote healthcare, innovation often means getting the basics right: reliable workflows, usable clinical systems, local navigation and face-to-face support before adding digital complexity.
- Technology selection should begin with frontline clinical usability. Systems that burden staff produce poor data, workarounds and weaker care regardless of their back-end reporting capability.
- Vulnerable communities should not be used to test immature systems. Mature clinical transcription and proven platforms offer more immediate value when supported by local training, governance and interoperability.
Data-Led Planning: Measuring Impact and Improving Community Outcomes
Kevin Harper, Visa Consulting & Analytics
Ian Hatton, City of Moreton Bay
- Fusing aggregated movement, spending and council information can help leaders understand growth earlier, test infrastructure scenarios and prioritise investment across a rapidly changing city.
- Data should improve the quality of debate rather than replace professional judgement. Planners, engineers and urban designers remain responsible for interpreting evidence in local context.
- The business case for data investment is strongest when it connects to organisation-wide and place-based outcomes, such as capital sequencing, visitor value and community access, rather than serving one department alone.
From AI Pilots to Governed Production
Scott Sheldon, DB Results
Lee Naidoo, OutSystems
- The gap between an attractive prototype and a production service lies in data quality, integration, security, operating cost, delegated authority and sustainable support.
- Before deployment, agencies should make approved assumptions and limits explicit through an operating contract that defines accountability, evidence, monitoring and acceptable conditions.
- Governance must continue after launch. Model, policy and environmental drift require observability, reassessment and a practical way to recall or disable an AI component without breaking the service.
Balancing Innovation and Risk in Modern Policing
Michelle Holland, Queensland Police Service
Laura Poidevin, Queensland Police Service
Cameron Allen, Queensland Police Service
- Risk appetite is an investment decision as well as a control question. Public organisations must weigh frontline value, opportunity cost and the consequences of failure within limited resources.
- Remote connectivity and AI trials create value when they solve a defined operational problem, have clear business ownership and can be supported within the wider technology environment.
- Layered controls, small-scale testing and frontline evidence provide a practical path to adoption while preserving human-led service design and organisational accountability.
The New Economics of AI: Managing AI-Driven IT Costs While Delivering Greater Public Value
Michael Coates, Apptio, an IBM company
- AI governance needs financial governance. Token charges are only part of the cost, alongside data readiness, training, operating-model change, infrastructure, assurance and ongoing support.
- Agencies should manage AI use cases as an investment portfolio, comparing total cost with service speed, capacity, quality, trust and other forms of public value.
- Finance, technology, data, security, risk and service leaders need a common view of consumption and benefits so they can expand, change or stop initiatives based on evidence.
Agile Innovation in Government: Is It Really Possible?
Lauren Poynting, Department of the Environment, Tourism, Science and Innovation, Queensland Government
Pia Andrews, Department of Children and Families, Northern Territory Government
Georgina Crundell, Queensland Audit Office
Vanessa Roarty, Digital Transformation Agency
Jeremy Janes, Queensland Rural and Industry Development Authority
- Institutional agility depends on delegated authority, clear outcomes and an operating model that includes finance, procurement, assurance and executive decision-making, not only delivery teams.
- Agile delivery still needs documentation and evidence, but assurance should test real risks and outcomes continuously rather than make document production the objective.
- Government already demonstrates rapid adaptation during emergencies. The leadership challenge is to build the same feedback loops, decision quality and cross-functional collaboration before a crisis occurs.
Published by
Help your peers
Share what you've learned with fellow public servants