In a recent interview with CXFocus’s Michael Gwilliam, at Ashton Media’s CX Retreat, Siân Howatson, Head of Customer Insights & Automation at Swyftx, discussed the evolving trends in CX, emphasising proactive and predictive customer experiences enabled by AI. She stressed the need for a balanced approach to AI implementation, ensuring it complements employee experience and drives meaningful business impact.
Michael Gwilliam (MG): Siân, what fundamental CX trends do you anticipate will reshape the Australian and international markets over the next 12 to 20 months?
Siân Howatson (SH): While there’s definitely been significant focus on AI, which is crucial to my role, I believe the fundamental shift is happening behind the scenes. The key trend I’m currently tracking and building towards is the delivery of proactive and predictive customer experiences.
While the concept of proactive experience isn’t new, we now have the technology to create these experiences at scale; something we lacked before. Previously, proactive efforts were often limited to one-to-one interactions, not the necessary one-to-many model. This shift is major. For example, instead of a customer contacting your support team about a missing deposit, you proactively send the customer a notification: ‘Hey, we’ve identified there’s a delay impacting your deposit; it should arrive in two hours.’
MG: Are you planning to tackle any of those today?
SH: We are building towards this over the next 12 months at Swyftx. Since enabling these features requires advanced technology to pinpoint the customer’s status, having a vision-first strategy is paramount before tackling the technical implementation. My philosophy is simple: don’t let the technology distract you; always start by defining the desired business outcomes.
In terms of trends, we’re seeing a significant shift toward internal team enablement, moving beyond the typical customer-facing applications of AI. We are focused on empowering our teams for smaller, day-to-day work with more robust AI copilots. While the first versions of these tools often fell short of expectations, the next iterations are far more powerful. Crucially, advancements in knowledge management—which powers the internal Gen AI—are now making these tools highly effective across the entire business, further feeding into our internal tools.
MG: What specific tools or resources are you leveraging to develop your copilots?
SH: We are primarily utilising off-the-shelf tools. Although, as a tech company, we have the capability to build our own—and are often tempted to—we evaluate every instance based on what makes the most sense from a cost and ROI perspective.
MG: How are you training your models, and what data sources are you utilising?
SH: For our internal copilots, the effectiveness is primarily content-driven; we need extremely high-quality content to make them useful for our team. Crucially, we maintain a strong internal feedback loop where the team can flag issues. While we have a dedicated team managing this, we heavily rely on that human feedback loop.
Overall, while adopting new ways of working requires a push from leadership, the team has been receptive. Our employees are generally early adopters of AI, so adopting new tools is seen less as a challenge and more as a slight adjustment to their workflow. The key lies in education to demonstrate that the new tools genuinely improve their important metrics. Once they grasp that concept, they quickly understand the value.
MG: Do you think most companies are currently overlooking the shift towards proactive customer experience?
SH: I believe most companies are engaging in proactivity, but the level of predictiveness is the key difference. True proactiveness involves anticipating all potential customer contact points and finding ways to make the contact seamless when it does occur. That is the level we are aiming for.
MG: Many CX leaders are struggling to translate customer insights to tangible business metrics or ROI. Do you have any frameworks or methodologies that can help people bridge the gap?
SH: It’s definitely a challenge for every team, including mine. We address this using a five-step CX value framework that directly links our work to the company’s bottom line. The process always starts with a business goal and connecting it to the required customer insights. We then determine the appropriate action and identify the specific KPIs it will influence, culminating in the dollar value or overall business impact. For example, a business goals could centre around increasing customer lifetime value or improving customer acquisition and onboarding.
Taking the onboarding example, my team first uses our omnichannel feedback program to identify the pain points within that critical flow (the Insight step). The Action is determined by a cross-functional team, who collaboratively define the steps needed to reduce friction. Our KPIs focus on core conversion metrics. Ultimately, the ROI is calculated by translating these improvements into financial metrics. For instance, if we reduce a 60% verification drop-off by X% and tie that to the cost of customer acquisition, we get a clear dollar value. This makes conversations about process adjustments much easier, as the ROI is quantified quickly.
MG: Regarding your AI agents, which I presume handle customer service enquiries, how have customers received them? Specifically, do you track data on their resolution rate?
Furthermore, what is your approach to transparency—do customers know they are interacting with AI, and does that knowledge affect their experience?
SH: The overall response is quite positive. We closely monitor our AI agent’s performance using an automated metric, our CX (Customer Experience) Score, which tracks customer satisfaction across all interactions. Currently, we maintain a 75% satisfaction CX Score, and this metric is increasing month-on-month. Customer feedback indicates this is a vast improvement over previous chatbot iterations and the automation we utilised prior to AI.
The experience is now far more conversational. Previously, interactions were constrained by ‘pick this button’ menus, but now customers can simply say what they want. This represents a major elevation in the experience. Making it clear that they are speaking to an AI agent and ensuring they can easily escalate to a human team member if necessary, is a critical part of our strategy.
MG: Siân, thank you so much for your time. The insights you shared today on proactive CX, AI enablement, and the CX value framework have been incredibly valuable.