The "Open Text" pattern has become the cornerstone of interactive AI design, fostering a dialogue between users and AI systems. This pattern is characterized by its simple interface that feels familiar, inviting the user to converse with the model underneath.

By using natural language, it doesn't take long for someone to get comfortable with the general interactivity. Where the pattern's limitations show is after the first few interactions, when someone doesn't know what to say next.

There's a false perception that simple means easy. When someone knows what they are looking for then this way of interacting with the model makes sense.  This could apply to use cases like a search portal, or customer support.

However, when someone reaches an open chat bar and doesn't know what they are looking for (content generation sites, ChatGPT, etc), it can lead them to feel crippled by the choices - the blank canvas effect.

On top of that, prompting skills are not widespread. Most users will not understand how to craft a prompt to get the result they have in their head.

Wayfinding patterns like ice breakers can help users get the conversation started. However, this pattern so far lacks affordances to help people construct better prompts that get them the outcomes they are looking for–the patterns are limited to hints and clues to get started. As result, users report feeling frustrated by the lack of consistency, predictability, or perceived quality in what is returned.

An open ended input pattern allows users to fully express themselves. They can use the words and framing that is more natural to them to construct a query. Open chat won't be going away, but we will likely see it evolve.

  • Templates can help users craft better prompts without having the full skillset
  • Nudges to improve your prompt can show users what "better" looks like
  • Putting filters and parameters at the users' fingertips can make this more complicate feature accessible

Think past the initial interaction. What's step two?

Details and variations

Considerations

Positives

Accessibility and Ease of Use
By leveraging natural language processing (NLP), the Open Text pattern makes powerful AI tools accessible to users without specialized knowledge. This democratizes access to technology, enabling a broader audience to benefit from AI advancements.

Flexible Interactions
The pattern supports a wide variety of user intents and queries, accommodating diverse needs and preferences. This flexibility enhances the user experience by providing personalized responses and solutions.

Enhanced User Engagement
The conversational nature of the Open Text pattern fosters an engaging and interactive user experience. It invites exploration and discovery, keeping users engaged and encouraging deeper exploration of the AI's capabilities.

Rapid Iteration and Feedback
Users can quickly iterate on their queries based on the AI's responses, leading to a dynamic interaction that feels more like a conversation with a human than an interaction with a machine. This immediate feedback loop helps users refine their queries and better understand the AI's capabilities and limitations.

Potential risks

Overload and Paralysis
The sheer openness of the interface can sometimes overwhelm users, especially those unfamiliar with the AI's capabilities or those who prefer more guidance. Without clear prompts or examples, users may struggle to initiate the conversation or articulate their needs effectively.

Misinterpretation and Ambiguity
Natural language is inherently ambiguous. Without the constraints of structured input, users might phrase queries in ways that the AI misinterprets, leading to unsatisfactory or irrelevant responses. This can frustrate users and erode trust in the system.

Privacy and Ethical Considerations
Given the open-ended nature of the interaction, users might share sensitive or personal information. This raises significant privacy and ethical concerns, necessitating robust data handling and privacy policies to protect user information.

Dependency on Natural Language Processing Accuracy
The effectiveness of the Open Text pattern heavily relies on the underlying NLP technology. Inaccuracies in understanding or generating responses can lead to user frustration, highlighting the importance of continuous improvement and refinement of the AI models.

Examples

The open chat that started it all - ChatGPT
Google took the ChatGPT interface, but they wait to reveal it until the user has submitted their first request
Anthropic's Claude chat bot looks familiar
Even when applied to specific interfaces like within Notion, the starting prompt is very open ended, relying on nudges and icebreakers to help users find what they need
In more specific contexts like Julius, the opening prompt is more specific
Jasper provides clear guardrails to the prompt - urging users to think of a writing task
Copy.ai relies on templates to get users started. Really, do they expect someone to search for *anything*?
When used within user flows the tool becomes more like a chat bot. Here, you can ask for feedback on your writing to Grammarly's chat
In a support setting, the pattern feels familiar. Note to "air answer" flair on the bot's response
No items found.