Reason Group's William Scheer articulates a counterintuitive thesis: AI implementation in government agencies can restore the human-centric customer service model that existed before centralised, efficiency-driven IT systems dismantled it. Rather than further automating away human interaction, Scheer argues that AI removes administrative friction—consolidating fragmented inboxes, automating routine governance tasks, and enabling self-service—thereby freeing frontline staff to reconnect with their organisation's mission and citizens' actual needs. This reframing positions AI not as a replacement for traditional service but as an enabler of it, allowing agencies to redistribute resources from bloated back offices to customer-facing roles. The practical example he cites—a client with 12 separate email inboxes now unified with AI-assisted knowledge management—illustrates the operational reality: AI handles the drudge work so humans can handle the relationship work.
The tension Scheer identifies, however, reveals why this vision remains largely unrealised in the public sector. Government agencies face genuine governance and accountability constraints that commercial CX teams rarely encounter. The "robodebt" precedent and concerns about AI scaling problems quickly have created institutional paranoia around algorithmic decision-making in core services, particularly where citizens' entitlements are at stake. This creates a critical question for CX leaders: how do you build trust in AI-assisted workflows when your stakeholders have witnessed high-profile failures? Scheer's incremental approach—rolling out ServiceNow HRSD with embedded AI capabilities to one client, monitoring for drift, adjusting in real time—suggests the answer lies not in proving AI's capability but in demonstrating governance rigour alongside it. For teams already operating in regulated environments or managing sensitive customer interactions, this raises a secondary consideration: is your monitoring and feedback loop infrastructure mature enough to catch and correct AI drift before it becomes a reputational or compliance issue?
Scheer's broader argument also exposes a structural assumption embedded in how many organisations have approached digital transformation over the past two decades. The shift from distributed, community-embedded service delivery (Centrelink offices in every town) to centralised systems was justified as efficiency optimisation, but it fundamentally altered the customer experience and employee engagement. If AI can genuinely reverse this trajectory by making distributed, human-led service economically viable again, it represents a philosophical shift in how CX teams should evaluate automation ROI—not purely through cost reduction metrics, but through mission alignment and frontline capacity. The question becomes whether your current CX platform and governance model can support this kind of strategic reorientation, or whether you're still locked into a cost-per-interaction framework that would penalise putting more people at the edge.
Why Reason Group CEO believes AI can bring back traditional customer service CRN Australia