AustralianSuper

AI agents will create a new tension inside organisations. Business teams will want more control. Technology teams will want more governance. And both will be right. Once an AI agent is live, the “business logic” inside it will keep changing. A policy changes. A process improves. A new exception appears. A regulator asks for evidence. A customer behaviour pattern emerges. A human reviewer spots a better way to handle a case. Business teams will not want to raise a ticket every time the agent needs a small instruction change. But technology teams cannot allow untested prompt edits, rule changes, tool changes, or workflow changes to go straight into production. Because with agents, even a small change can create a new version of behaviour. So the question is not: “Who owns the agent?” The better question is: “What parts of the agent should business own, what parts should technology own, and what must be jointly governed?” I think the operating model needs to separate the agent into clear components: Business owns: Outcomes, policies, rules, exceptions, approval thresholds, tone of judgement. Technology owns: Architecture, tools, integrations, data access, security, observability, release controls. Jointly owned: Evaluations, guardrails, version approval, failure review, escalation paths. The mistake is putting everything into one black-box prompt and calling it an agent. That creates either chaos or bottlenecks. The better model is a governed agent design: Business can propose and maintain business logic. Technology provides the control layer. Every change is tested, approved, versioned, monitored, and reversible. Not a free-for-all. Not an IT ticket queue for every small change. A controlled change loop. AI agents do not just need a delivery model. They need an ownership model. Because once agents start doing real work, the most important question becomes: Who is allowed to change the agent’s behaviour, under what controls, and with what evidence? How are your teams thinking about agent ownership: business-led, technology-led, or jointly governed? #AgenticAI #AIAgents #ProductManagement #DigitalTransformation #SoftwareDevelopment

We have launched our educational workshops for Australians over 65! A huge thank you to the Gordon Pymble Uniting Church for hosting our very first session on “Financial Safety: Steps to Safeguard Your Financial Independence.” During the workshop, we covered simple, practical steps to detect scams and financial abuse early. The focus was on empowering older Australians with knowledge they can apply on their own to protect their financial independence. 🛡️ I also had the opportunity to demonstrate the My Lasting Plan prototype. Our tools are specifically designed to make these practical steps easy to apply. The feedback we received will be instrumental for our ongoing product development.📈 My Lasting Plan is a local social enterprise. We view financial safety as a fundamental pillar of holistic retirement wellbeing. Our mission is to empower older Australians to age with dignity and confidence. 🌍 If your organisation is based in Sydney and is interested in hosting an educational workshop for your community, please don’t hesitate to reach out. Let's solve this social problem together! 🤝 #SocialEnterprise #FinancialSafety #AgeTech #AgeingWell #TechForGood #MyLastingPlan #CommunitySupport #UnitingChurch

🚀 We’re hiring – Senior Technical Security Specialist We’re looking for a Senior Technical Security Specialist to join my security operations team. This is a hands‑on, impact driven role sitting at the intersection of security engineering and cyber resilience, focused on strengthening detection and response at enterprise scale. It’s a great fit for someone who enjoys solving real problems and influencing meaningful outcomes. https://lnkd.in/gpYdFiX4

I recently used an AI agent (Hermes by Nous Research ) to automate a tedious process for a small business. The gain? 75% reduction in effort and better quality of client biomentric data analysis with AI powered insights based on real world knowledge. And the pattern applies to just about any workflow. Read more below...

Have you ever wondered what it takes to build something professional and high-quality with AI? Or why you're not getting the kinds of results that you thought you'd get by asking your agent to "build me a SaaS app to do XYZ"? Here is my take on that question and a professional engineering team grade harness that everyone can use without any training. https://lnkd.in/gufRPKqN

View my verified achievement from The Institute of Internal Auditors - Australia.