For an organisation that connects millions of Australians every day, the shift toward artificial intelligence is about amplifying the human element.
At the centre of this transformation is Michelle Francis, General Manager of Customer Success at Australia Post. Managing an ecosystem that serves millions of senders and receivers, Michelle’s mission is clear – modernise the customer experience while maintaining the trust and empathy that Australians expect from the brand.
In this interview, Michelle pulls back the curtain on how Australia Post is navigating the “AI Paradox.” From reducing tech-stress among frontline staff to designing AI in ways that strengthen human judgement and capability, she shares the practical blueprint for a successful digital evolution.
NOTE: Michelle will be presenting at the upcoming Contact Centre Symposium in the Hunter Valley organised by Ashton Media.
Mark Atterby (MA): Please provide a background to your career and job role.
Michelle Francis (MF): I lead Customer Success at Australia Post, which essentially is ensuring we deliver great customer service for our receiver and our sender customers. A big part of my role right now is helping Australia Post modernise how we support our customers. We talk about it as upgrading the experience that Australia Post delivers: more convenience, more choice, faster resolution, and a more connected experience across digital and human channels.
MA: Please provide a description of the team that you lead and its mission.
MF: The Customer Success team is here to create outstanding experiences for every sender and receiver. We bring together the operational leadership of frontline customer service teams that assist customers with everything from “where’s my parcel” queries, through to onboarding and portal support for our sender customers. Customer Success also leads the capability-building pieces: knowledge, coaching, insights, continuous improvement, and the tools that remove friction for both customers and our team members.
Our mission is to resolve customer needs with clarity and care, and keep upgrading the experience. It’s about improving the everyday service experience customers have with Australia Post, and building the tech and support our people need behind the scenes to bring that level of
MA: How does AI-driven transformation in the contact centre specifically support the customer goals of Australia Post?
MF: For our customers, the goal is straightforward: get the right answer quickly, with confidence, and be treated like a person, not a ticket number. We see AI as a partner to our team members in providing that level of service – it supports by resolving simple issues quickly, lifting consistency in answers, and freeing up our people to focus on the moments that really need human judgement and empathy.
Our commitment to enhancing convenience and choice for customers is central to the broader service upgrades we’re delivering at Australia Post. This approach is reflected in every aspect of our customer experience strategy, underpinning our promise to make sending and receiving parcels as straightforward and flexible as possible.
One practical example of this commitment is our investment in Out-of-Home solutions, such as Parcel Lockers. By providing these options, we give customers greater flexibility and control over their deliveries, allowing them to collect items at a time and place that suits their lifestyle. This focus on convenience addresses the growing need for adaptable services and ensures that customers have more choices when interacting with Australia Post.
MA: Can you mention any pilots that have been undertaken and how they are helping to inform your broader AI strategy?
MF: Yes, we’ve taken a very practical approach: start small, test in the real environment, and scale only what genuinely improves the experience for customers and for our team members. Our pilots have included AI interaction summaries that reduce the admin load so our team members can spend more time helping customers, and triaging support so that customers get to the right help path sooner.
What we learn in pilots isn’t just “does the tech work?” – it’s what changes are needed in coaching, knowledge management and workflow so it works reliably at scale. That’s what shapes our broader AI strategy: build the foundations, then expand.
MA: AI should “replicate the best of human experience.” How do you ensure the technology doesn’t strip away the human touch?
MF: The goal is for humans to partner with the tech to create the best possible experience. So our approach is: AI should handle the repetitive and time-consuming parts, and our people should own the relationship and the decision-making. Customers can tell when something feels robotic, so we’re intentional about keeping the language natural, respectful and consistent with how Australia Post shows up.
MA: What specific training and cultural programs have been rolled out to help contact centre agents become proficient with AI tools? Which tools have been adopted?
MF: We’ve treated this as a capability and culture shift, not just a tech rollout. So we’re focusing on upskilling and clear expectations, because AI changes the shape of the job and we want our people to feel confident in that change.
In terms of training, we’ve focused on three things:
1. How to use AI responsibly
2. How to keep responses customer-friendly and on-brand
3. How to escalate and apply judgement in complex or sensitive cases
MA: How are frontline agents involved in the design of the AI tools they use? Can you share an instance where agent feedback changed how an AI feature was implemented?
MF: Frontline involvement is non-negotiable for us – our team members say these tools are “designed by us, for us”. If the tool doesn’t work in the flow of a real customer conversation, it won’t stick. So we involve our team members from the start: they tell us what they need, we test prototypes with them, we run feedback loops, and we measure success based on whether it genuinely saves time or improves clarity for the customer.
MA: AI can be a source of ‘tech-stress’ for some. How has Australia Post managed the psychological transition for staff?
MF: That’s a really important point – tech-stress is real, especially when people worry about job security or feel they’re expected to learn everything overnight. We’ve managed the transition by pacing change sensibly, building confidence through training and peer support, and reinforcing that humans still own the customer experience. When people feel safe to ask questions, adoption and wellbeing both improve.
MA: How does Australia Post ensure that the AI-created content used by agents remains on-brand and, more importantly, accurate for the customer?
MF: We take governance seriously. Accuracy is everything in customer service – if the answer is wrong, you don’t just lose time, you lose trust. That’s why we use approved key messages and a messaging governance approach to keep content consistent and on-brand.
In practice, we combine guardrails (what the tool is allowed to draw from), clear standards for tone and language, and human accountability. And we keep improving the underlying knowledge, because AI is only as reliable as the information it’s grounded in.
MA: What are the 2-3 most critical skills a contact centre agent needs today compared to five years ago?
MF: First, digital fluency – not just using systems, but confidently moving between channels, tools and information sources. Second, critical thinking and judgement – knowing when something doesn’t look right, when to verify, and when to escalate. And third, human connection: empathy, de-escalation and communicating clearly, especially when a customer is frustrated or anxious.
MA: What is the one lesson Australia Post has learned about the “People” pillar of AI transformation that you would share with other Australian organisations?
MF: Be honest that this is a journey – the goal is progress, not perfection. We’ve also learnt to start with the workflow and the frontline reality, to co-design with the people who will use it, and to invest in capability uplift early. When you do that, AI genuinely becomes an upgrade for customers and employees, not another layer of complexity.
