The Limitation and Discomfort of Using a Chatbot
Customers often value genuine interactions, especially during important buying decisions. AI interactions that closely mimic human behavior while still displaying noticeable limitations can sometimes create discomfort or skepticism. This can become particularly apparent when an AI interaction feels human-like in some ways but still lacks the nuances of authentic human conversation.
As explained by Trone, “Once AI gets so realistic that it’s hard to distinguish between robot and human, brands may start finding that the uncanny valley is not only a problem with physical appearance… but also a problem with language, too.” This idea applies the uncanny valley concept to language and conversational AI, where interactions that feel almost, but not completely, human may create discomfort for some users.
The uncanny valley concept, originally coined by Japanese roboticist Masahiro Mori in 1970, describes the idea that people’s affinity toward an artificial entity may decline when it becomes highly human-like without appearing completely natural. In chatbot conversations, similar discomfort may occur when systems:
- Use overly familiar language that feels forced or inauthentic
- Fail to grasp emotional context or subtle conversational cues
- Respond with technically correct but contextually inappropriate answers
- Switch awkwardly between scripted responses and generated content
- Miss the emotional weight behind customer questions or concerns
An example shared by Rebecca Emily Darling, Director of Storytelling at Huxly, illustrates this type of experience. When interacting with an AI chatbot, she pointed out that the bot’s avatar was wearing beige clothing. According to her account, the bot repeatedly denied this, responding, “I don’t wear beige — I wear fierce! Beige is for basic girls — I’m a queen!” The example illustrates how an obviously incorrect but highly human-like response can make an AI interaction feel strange or uncomfortable.
For businesses, these kinds of interactions can affect the customer experience. As discussed by CMSWire, highly lifelike virtual assistants or AI characters can potentially create discomfort and negative associations when the experience does not feel natural.h
Some customers may experience discomfort, uncertainty, or frustration and decide to end the interaction, including at moments when they may be actively considering a purchase.
The Knock Knock App Difference
Knock Knock App addresses this challenge by combining AI-powered engagement with opportunities for authentic human interaction. Our platform uses GPT-4-driven chatbots for initial lead qualification, helping businesses handle appropriate initial interactions efficiently. The key difference lies in our human-AI collaboration model: when appropriate, qualified customers can be transitioned into real-time video or audio conversations with actual sales representatives who are available to continue the conversation.
This approach also reflects an important consideration raised by Tony Fernandes, CEO of UserExperience.AI: “If a brand does not disclose that the experience is artificial, people will eventually find out and feel fooled. Leaving people with the feeling that you tried to fool them is never a positive brand attribute.”
Knock Knock doesn’t need AI to pass as human. Instead, AI can be used for tasks such as information gathering, answering supported questions, and initial qualification, while humans remain available for conversations where personal interaction, expertise, or judgment is valuable.
Why It Works
Humanized Interaction
Knock Knock allows businesses to transition from AI interactions to human conversations when appropriate. This approach acknowledges the limitations of AI while leveraging its strengths, creating a customer journey that can combine efficiency with authentic human interaction.
The transition can happen when a prospect shows relevant interest signals or when the conversation reaches a point where human involvement would be useful. Rather than expecting AI to imitate a human indefinitely, businesses can use escalation rules and human takeover to provide another option when the conversation requires it.
Instant Engagement
When a team member is available, connecting through video can reduce the need to move every interested visitor into a separately scheduled meeting. For some customers and sales situations, this can provide a more direct way to ask questions, see demonstrations, and interact with a real person.
As noted in discussions of the uncanny valley effect, successfully navigating highly human-like digital interactions can contribute to a more comfortable user experience. Knock Knock approaches this by not requiring AI to imitate every aspect of human conversation, but instead creating a seamless handoff to human representatives when human involvement is appropriate.
Personalized Approach
The AI captures detailed visitor behavior, providing human agents with available context when they enter conversations. This can help representatives understand relevant visitor activity and provide more personalized interactions without requiring the customer to repeat everything from the beginning.
Bridging the Valley
Knock Knock App is designed to balance AI efficiency with human interaction. Customers can benefit from the speed and convenience of AI while businesses retain the ability to introduce human involvement when it is useful.
The key idea is that businesses don’t necessarily need to make AI increasingly indistinguishable from humans. Instead, AI and human representatives can each be used where their respective capabilities are most useful.
As Rebecca Emily Darling notes, “Conversation is the future of marketing—and the art of conversation is the exact thing that AIs are still struggling to grasp.” Knock Knock embraces this approach by using technology to support rather than eliminate the human connection that can matter during sales conversations.
Real Results
Rather than assuming that every business will achieve the same conversion improvement, the impact of combining AI with human engagement should be measured using each company’s own customer journey and performance data.
Relevant outcomes to monitor can include:
- Changes in visitor-to-conversation rates
- Customer drop-off at important decision points
- Customer satisfaction with AI and human interactions
- Changes in sales-cycle length
- Conversion rates and average deal values
These results can vary based on traffic quality, audience, offer, implementation, AI configuration, sales process, and team availability. Comparing these metrics with your existing baseline provides a more reliable way to determine whether the approach is improving the customer experience.
Final Thought
In a marketplace filled with AI-driven solutions, businesses should recognize both the capabilities and limitations of artificial interactions. The uncanny valley is an established concept in human responses to highly human-like artificial entities, and similar ideas are increasingly discussed in relation to conversational AI.
That doesn’t mean every human-like chatbot will make customers uncomfortable, or that AI interactions inherently reduce trust. Customer reactions depend on factors such as transparency, accuracy, context, design, expectations, and the quality of the interaction.
Knock Knock App combines AI-powered engagement with opportunities for human interaction, allowing businesses to use automation where it is useful while keeping genuine human conversation available when it matters.
Rather than assuming this approach will produce a fixed conversion increase, businesses can evaluate its impact using their own engagement, customer experience, lead, and conversion data.
Anna Hyatt
Co-Founder | Knock Knock App
AI × Human Touch = Happier Clients, More Revenue
LinkedIn Profile
