A disclosure banner stating "You're chatting with our AI assistant" has become standard practice across contact centres, yet this surface-level transparency masks a fundamental problem: customers cannot meaningfully understand AI's role in their interaction based on a label alone. The issue intensifies as AI embeds itself throughout customer journeys rather than remaining confined to identifiable chatbots. A single interaction might involve an AI agent handling initial contact, a human agent deploying an AI copilot, and another AI system executing backend actions—yet a generic disclosure banner fails to capture this complexity. The real challenge for CX teams is distinguishing between compliance theatre and genuine transparency. Srinivas Chippagiri from Salesforce articulates this precisely: "The requirement is met; the person is no better informed." Disclosure strategies that treat AI notification as a checkbox exercise across all channels—applying identical banners to chat, voice, and asynchronous messaging—fundamentally misunderstand how customers process information. Voice interactions demand early, clear verbal disclosure; chat benefits from persistent visual indicators; email and messaging require different timing considerations entirely. The question becomes whether your current disclosure approach actually prevents customers from misunderstanding who or what is making decisions on their behalf, or whether it merely protects your organisation from liability.
Trust cannot be engineered through disclosure alone; it requires genuine agency. Research by Dr. Eva Wolf demonstrates that AI disclosure paradoxically increases scepticism in high-stakes interactions, yet builds trust when customers perceive human oversight. This creates a critical tension for teams managing complex workflows: disclosing AI involvement without offering a functional path to human escalation actively erodes trust. Chippagiri's observation that "disclosing that AI is involved does nothing if the escape hatch does not actually open" reflects a widespread failure—systems that promise human handoffs but return customers to the same bot, or force information repetition, transform transparency into frustration. For CX leaders implementing AI-assisted tools like Salesforce's Agentforce or similar platforms, this means your disclosure strategy must account for materiality: when does AI involvement genuinely shape outcomes or customer decisions? Behind-the-scenes uses such as summarisation, routing, or spell-checking warrant different treatment than AI making eligibility, pricing, or denial decisions. The practical implication is layered disclosure—plain-language explanations of AI's role with optional deeper detail—rather than technical audits customers cannot act upon.
The strategic shift required is moving from "Did we disclose AI?" to "What should customers reasonably understand about this specific interaction type?" This demands mapping disclosure practices against interaction stakes, AI involvement levels, and communication channels simultaneously. Hybrid interactions present the most acute challenge: customers need awareness when AI materially influences outcomes without drowning in technical detail about system architecture. For teams already running multi-channel operations with embedded AI, this means auditing whether your current disclosure practices actually inform customers or merely satisfy compliance requirements. The goal is preventing false understanding about who controls decisions and actions—whether human or AI—not creating exhaustive technical transparency that obscures rather than clarifies.
The AI disclosure banner is here for customer service, but trust will take more work. What should meaningful disclosure for interactions with enterprise AI systems look like in contact centers?
Don't confuse a disclosure banner for credibility No Jitter