DXC Technology and SAP panel discussion | August 1, 2025
Agentic AI in the public sector
At the recent SAP Government Innovation Day in Canberra, DXC Technology hosted a compelling panel discussion on the transformative potential of Agentic AI in the public sector. The session, moderated by Denise Lucero, SAP Practice Leader for Australia and New Zealand at DXC, featured insights from Dr Chris Nøkkentved, Global CTO for SAP and Enterprise Applications at DXC, and Ryan van Leent, SAP’s Global Vice President for Public Services. Together, they explored how AI, data, and enterprise applications are converging to drive agility, compliance, and innovation in public finance.
Understanding Agentic AI vs. Generative AI
Chris Nøkkentved opened the discussion by clarifying the distinction between Generative AI (Gen AI) and Agentic AI. While Gen AI is often associated with large language models and neural networks that simulate human-like responses, Agentic AI represents a more collaborative and goal-oriented approach. It involves networks of AI agents working together to solve complex problems - mirroring the collaborative nature of public sector operations such as case management and service delivery.
Agentic AI integrates various AI models, including machine learning for pattern recognition, expert systems for compliance, and vision models for asset inspection. These agents are designed to support, not replace, human decision-makers - particularly in high-stakes areas like finance approvals and policy enforcement. The emphasis is on augmenting human capabilities, distributing workloads, and accelerating productivity through intelligent orchestration. Advanced agents focus on supporting issue resolution in an “explainable” manner which leverages different models merging knowledge networks with GenAI constructs.
Modernising without major transformation
A key concern for public sector organisations is how to adopt AI without undergoing disruptive system overhauls. Dr Chris emphasised that SAP’s Business Technology Platform (BTP) enables integration with legacy systems, allowing agencies to gradually adopt AI capabilities. This includes advanced user interfaces, robotic process automation (RPA), and decision-support models that can be layered onto existing infrastructure.
The introduction of SAP’s Business Data Cloud, in collaboration with Databricks, further enhances this capability. It allows seamless data collaboration between SAP’s ERP systems and big data platforms, enabling faster, more cost-effective processing without the need to move data externally. This opens new possibilities for managing complex financial and procurement cases, as seen in real-world implementations across New Zealand and Europe.
AI as the new user interface
Ryan van Leent opened with a compelling example from City of Antibes in the south of France, where AI is used for public budget optimisation and visualisation, and linking of budgetary operations to sustainable development goals. Specifically, they’re using AI to align 6,000 budget credit lines with the 17 Sustainable Development Goals (SDGs) and six pillars of their green budget initiative. This results in 138,000 AI-driven decisions - far beyond human capacity - demonstrating the power of AI in public finance optimisation.
Ryan explained the industry shift from custom AI projects to adoption of AI capabilities that are embedded in enterprise applications. He shared examples from SAP’s roadmap in which AI capabilities are being embedded directly within its business suite, and how SAP is prioritising features that offer the greatest business impact. Interestingly, it’s often the simpler tools - like AI-assisted search, filtering, and summarisation - that provide the most value. These are being delivered through SAP’s Joule copilot, which acts as an intelligent intermediary between users and applications.
At SAP Sapphire (SAP’s annual conference held in Orlando, Florida, from May 19-21, 2025) demonstrations showcased how users now interact with enterprise systems via natural language prompts to Joule, rather than traditional UI navigation. This paradigm shift - “AI is the new UI” - has profound implications for both users and developers, as it redefines how enterprise software is designed and experienced.
Transparency and responsible AI
Addressing concerns around data governance, explainability, and accountability, especially in the public sector, Ryan stressed the importance of transparency and reasoning in AI processes. This is especially the case in multi-agent systems, where an early error can cascade through the workflow. To mitigate this, SAP provides an “AI Analysis” tab, which provides transparency in the agentic process and the reasoning behind AI recommendations.
This transparency empowers human users to remain in control, making informed decisions rather than acting as passive overseers. It also aligns with principles of responsible AI, ensuring that public sector implementations are ethical, auditable, and trustworthy.
From proof to scale: The AI adoption journey
Dr Chris outlined a pragmatic approach to AI adoption in the public sector, emphasising the value of starting small. He cited the City of Brussels as an example, where AI is used to enhance citizen services, provide legislative guidance, and resolve disputes. These early-stage projects serve as learning platforms, gradually building organisational confidence and capability.
The journey typically begins with expert systems and RPA, progresses to machine learning for issue resolution, and eventually incorporates advanced AI agents for real-time interaction. Importantly, SAP and DXC are working to pre-build models that customers can adopt and customise, reducing the burden of developing complex AI solutions from scratch.
Conclusion
The panel concluded with a shared vision: AI should simplify - not complicate - public sector operations. Whether through enhanced decisionmaking, streamlined procurement, or improved citizen engagement, Agentic AI offers a path to smarter, more responsive government. By starting small, leveraging existing systems, and prioritising transparency, public sector organisations can harness AI to drive meaningful change - without the need for massive transformation
Q & A summary
1. What is the difference between Generative AI and Agentic AI?
Generative AI focuses on creating content using large language models, while Agentic AI involves multiple AI agents working collaboratively toward a shared goal, integrating various models for decision-making, compliance, and automation.
2. Can public sector agencies adopt AI without replacing legacy systems?
Yes. SAP’s Business Technology Platform allows agencies to integrate AI capabilities with existing systems, enabling gradual adoption through advanced interfaces, RPA, and data collaboration tools like Databricks.
3. How is AI changing the way users interact with enterprise applications?
AI is becoming the new user interface. Tools like SAP Joule allow users to interact with systems via natural language, automating tasks and simplifying navigation across complex applications.
4. How does SAP ensure transparency and accountability in AI processes?
SAP provides transparency into agentic processes and AI reasoning, ensuring that human users remain informed and in control, which is critical for responsible AI in the public sector.
5. What’s the best way for public sector organisations to begin their AI journey?
Start with small, manageable projects such as expert systems or RPA. Use these to build internal knowledge and confidence, then scale up to more advanced AI solutions using pre-built models and platforms provided by SAP and DXC
Further reading - strategic perspective from Dr Chris Nøkkentved
For more from DXC’s Dr Chris Nøkkentved’s, the article “SAP + Databricks: Putting Data and AI to Work” highlights the strategic integration of SAP’s Business Data Cloud with Databricks. This integration enables unified access to structured and unstructured data without duplication, preserving SAP data context and enhancing AI-readiness.
The article outlines how this synergy supports use cases such as supply chain optimisation, inventory management, and profit forecasting. It also emphasises the importance of cloud flexibility, allowing public sector organisations to modernise at their own pace while maintaining data governance and compliance.