Shana Simmons, Zendesk's Chief Legal Officer, has articulated a fundamental challenge facing CX organizations deploying AI at scale: traditional governance models built around centralized legal review and rule-based compliance cannot keep pace with distributed decision-making and accelerating delivery cycles. Her argument moves beyond the familiar rhetoric of "guardrails" and "human-in-the-loop" to propose principles-based governance as an operating model intervention—one that embeds legal and compliance judgement into the flow of work rather than treating it as a gate at the end. The core insight is deceptively simple: universal principles scale better than context-specific rules because they can be absorbed into team culture and decision-making from the outset, whereas traditional compliance training and checkbox exercises create the illusion of control without changing behaviour. For CX teams already managing complex AI-assisted workflows—whether through Zendesk's own AI features or integrated third-party agents—this distinction matters acutely: if governance remains siloed in Legal, the pressure to maintain delivery velocity will either create backlogs that undermine the speed advantage AI promises, or push teams to route around formal review processes entirely.
The operating model shift Simmons describes requires product teams to internalize governance as part of their value proposition rather than an external constraint imposed after the fact. This means embedding principles around privacy, transparency, accuracy and accountability into how support teams design workflows, configure agents, and handle customer data—not as compliance theatre but as foundational product decisions. The leverage point is customer demand: enterprise customers increasingly scrutinize how vendors build AI systems, making governance a competitive differentiator rather than a cost centre. For support leaders and CX consultants, this reframes the relationship between your teams and Legal from adversarial (Legal slowing you down) to collaborative (Legal helping you understand what responsible delivery looks like in your context). The question becomes whether your organization has created the conditions for this shift—whether product and support teams have absorbed enough governance thinking to consider compliance implications before Legal arrives, or whether governance remains something that happens to you rather than something you do.
Simmons' federated governance model suggests that as AI distributes the ability to build and automate across your organization—from support agents configuring workflows to analysts designing customer journeys—the ability to judge risk and compliance must distribute alongside it. This does not mean governance becomes less rigorous; it means governance becomes continuous rather than episodic, embedded in tools and accessible support rather than concentrated in formal review gates. For CX professionals managing teams that increasingly use AI to make decisions affecting customers, this is not optional refinement but structural necessity. The alternative is either accepting that everything queues behind Legal (defeating the point of AI acceleration) or tacitly accepting that teams will work around centralized governance to protect delivery timelines. Neither outcome serves customers or the organization.
Zendesk’s Shana Simmons on why AI governance cannot live in Legal alone Diginomica
Zendesk CLO Shana Simmons: Empathy is the new superpower for AI leaders MSN