What Should a Chatbot Say When It Cannot Help and Needs to Escalate?

Chatbots have become essential tools for delivering fast, efficient customer support—especially within SaaS businesses Find out more serving European Union users. However, no AI assistant can cover every situation perfectly. When a chatbot encounters a question beyond its capabilities, the way it communicates this limitation and escalates to a human agent is critical. Done well, it can build trust, reduce customer frustration, and comply with evolving legal standards such as the EU AI Act.

The Importance of Clear Handoff Language and Expectation Setting

When a chatbot cannot adequately address a user’s query, the phrase “escalate to human” is more than a technical handoff. It represents a pivotal moment in the customer journey. Poorly handled escalation can lead to confusion, dissatisfaction, or worse, loss of trust. On the other hand, clear handoff language and transparent expectation setting help users understand what’s happening and what to expect next.

Take, for example, Coruzant Technologies, a company that has refined their customer support chat interfaces to openly disclose when users are being transferred to a human agent. Their UX writing avoids generic phrases like “connecting you to support” and instead says something like, “I’m here to help, but this question is best handled by one of our specialists. I’m connecting you now, and they will join shortly to assist.” This kind of messaging reassures users and manages expectations clearly.

Legal Context: The EU AI Act and Article 50 Transparency Requirements

Beyond UX best practices, chatbot escalation must comply with legal regulations. The EU AI Act, recently proposed to regulate artificial intelligence systems, includes Article 50, which establishes important transparency obligations for AI providers.

What Does Article 50 Require?

The essence of Article 50 is that users must be informed that they are interacting with an AI system, and they must be made aware of the AI system’s capabilities and limitations. When the AI system cannot help and escalates to a human, this transition must also be communicated transparently.

For instance, a chatbot should not leave users wondering if they are still communicating with an AI or a human. The AI must “describe itself” clearly and set accurate expectations, including the reason for escalation and what the human agent will do next.

Provider vs Deployer Responsibility

One complexity in compliance to Article 50 involves the division of responsibility between the AI system provider (the company developing the chatbot technology) and the deployer (the company integrating and deploying the chatbot in their service). For example, Coruzant Technologies as a SaaS provider may supply the chatbot, but a client deploying the bot is responsible for the transparency disclosures shown to EU users.

Both parties must collaborate to ensure disclosures about AI limitations and handoffs are prompt, clear, and compliant, especially since non-EU companies deploying chatbots for EU users are subject to extraterritorial reach of the EU AI Act—in other words, the regulation applies regardless of where the company is located, reinforcing the global importance of clear chatbot escalation protocols.

Accessibility Considerations: Supporting Users with Screen Readers and Voice Interfaces

Effective chatbot escalation messages must also be accessible to all users, including those relying on assistive technologies such as screen readers and voice product interfaces. Legal experts like the European Data Protection Supervisor (EDPS) emphasize that inclusivity is a critical component of user transparency and consent.

For instance, messages stating, “I’m escalating this to a human now” may seem straightforward when reading visually, but to a screen reader user, or when conveyed over voice interfaces, phrasing impacts how understandable and natural that notification is.

Screen reader compatibility: Use simple, unambiguous phrasing and avoid jargon. For example, instead of saying “This conversation will now be transferred to a Tier 2 specialist,” say “I’m connecting you to a human support agent who can better assist you.”
Voice interface tone: When the chatbot’s voice interface is the primary communication channel, it should read escalation messages naturally, avoiding robotic or disjointed language that can confuse or frustrate users.

Reading messages out loud during development helps UX writers and product teams catch awkward phrasing and ensure smooth interactions.

When to Disclose: Timing of the First-Interaction Disclosure

Another key question is when users should be informed they are interacting with an AI system, and when the system cannot help. The European Commission’s guidance suggests transparency disclosures should occur at the first meaningful interaction—which means early enough so users know they are dealing with an AI assistant, not buried somewhere later in the conversation or hidden in terms of service.

For instance, Coruzant Technologies includes an introductory chatbot message such as:

"Hello! I’m CoruzantBot, an AI-powered assistant here to help fix issues quickly. If there’s a question I can’t answer, I’ll connect you directly to a human expert."

This upfront statement accomplishes multiple requirements:

Identifies the AI nature of the assistant.
Sets expectations about the AI’s abilities and boundaries.
Prepares the user for escalation if needed.

Users appreciate this transparency because they know going in what kind of experience to expect.

Examples of Effective Chatbot Escalation Messaging

Here are several examples illustrating best practices for chatbot messages Perplexity that clearly explain escalation to a human agent:

Message Feature Example Text Why It Works Clear identification of chatbot nature "Hi, I’m your virtual assistant here to help you today." Sets AI identity upfront, per EU AI Act Article 50 Limit acknowledgement "I’m sorry, I don’t have enough information on that." Conveys system limitation transparently Immediate escalation offer "Let me connect you to a human specialist to assist you further." Sets clear expectation for next step Accessibility-friendly phrasing "I’m transferring you to a person who can answer your question better. Please wait a moment." Simple language, easy for screen readers to process Extrajurisdictional clarity "This AI assistant follows EU data protection standards to ensure your privacy." Reassures users on cross-border compliance

Collaborating Across Legal, Accessibility, and Product Teams

Building chatbot escalation messages that satisfy legal transparency, user experience clarity, and accessibility requires cross-functional teamwork.

At companies like Coruzant Technologies, product managers, UX writers, legal counsel, and accessibility experts come together to create launch checklists that include:

Review of all escalation language for clarity and compliance
Testing messages with assistive technology such as screen readers
Documenting disclosure timing aligned with EU guidelines
Validating voice interface message tone and pacing

These collaborations minimize the risk of support tickets caused by unclear UI or legal violations due to insufficient transparency disclosure.

Conclusion

When a chatbot cannot help and needs to escalate to a human, precise, transparent, and accessible communication is not optional—it’s essential. Companies must strike a balance between meeting user experience expectations and complying with the EU AI Act’s Article 50 transparency rules, especially given their extraterritorial impact.

Start by informing users upfront that they’re interacting with an AI, clearly stating the bot’s limits, and setting expectations on escalation timing and process. Craft handoff language that is simple, inclusive, and natural—especially for voice and screen reader users. Collaborate closely with legal and accessibility teams to refine and verify the experience.

By following these guidelines, SaaS providers operating in or serving the European Union can build chatbot experiences that not only delight users but also meet strict regulatory standards—enhancing trust and customer satisfaction in every interaction.