The UX battle of Generative AI in text
AI technology has entered its teenage years, proving its potential but still a bit awkward in terms of incorporation into workflows and products. In recent months, text applications have incorporated Generative AI-powered GPT technology to augment UX flows. In this article, we will explore different ways to incorporate the power of AI text generation in your product, from auto-complete to interactive chatbots and shortcuts.
Auto-complete: streamlining workflows
Auto-complete is a familiar tool that increases typing speed by suggesting the next word or phrase to the user. It isn’t anything new, but it is a good starting point to compare the other UX techniques on this list. The main advantage of auto-complete is that it streamlines workflows and saves time, as it helps to finish the current word or suggests the next word to type. The fact that we’ve gotten so used to auto-complete is a testament to its added value in UX.
Text Completion: Sparking Creativity
So, what comes next? Let’s talk about text completion. Text completion goes one step further, and leverages the powers of Generative AI by creating entirely new content based on the existing context. It allows users to explore ideas and thoughts they might not have considered otherwise, breaking free from their usual thought patterns and enhancing the overall quality of their output. LEX, a lightweight text editor, is an example of text completion that strikes the right balance between automation and human touch.
As we dive deeper into the world of Generative AI, it becomes evident that the potential for creativity and enhanced user experience is immense. The key is to strike the right balance between automation and the human touch while ensuring that the technology supports and empowers users rather than replacing them.
Chat: collaborative co-creation
Let’s move away from this one-sided conversation and start to collaborate in the creative process. By asking questions and working together, both users and AI can co-create content and refine ideas, leading to a more collaborative approach to the writing process.
This interactive method is especially interesting for ‘AI beginners’. The iterative approach makes it less dependent on prompt engineering and more forgiving to use. Users can benefit from the AI's ability to understand and respond to their queries, while the AI can learn from the user's input, thereby improving its suggestions over the conversation.
The sky really is the limit when it comes to the number of possible conversations you can have. To really experience this, we would recommend trying it out yourself. Naturally, you can experiment with ChatGPT. But we took it one step further. At In The Pocket, we created ChatITP, a Slackbot which integrates into our Slack environment and adapts to the context and way of working of In The Pocket. It is a low-entry bar for all employees to leverage the power of chat and allows us to experiment with these technologies.
Shortcuts: Streamlining the process
As we move forward in our exploration of Generative AI, it's clear that the flexibility of chat-based interfaces comes with some inefficiencies. Having a conversation with AI can be time-consuming. And although conversations are a natural way to communicate, it does take you out of your working flow. So, what's the alternative to overcome these limitations?
Enter Shortcuts (trademark pending) – a more efficient way to harness the power of Generative AI. By introducing specific, repeatable actions based on context, these shortcuts enable users to bypass the need for prompt engineering and directly access the desired information or assistance.
Under the hood, we are still using the same language-based technologies, prompt engineering and fine-tuning, but we integrate the complexity and need for conversations away behind some nice UI. This allows for a more streamlined user experience and reduces the time spent on repetitive tasks.
Moving from auto-complete to text completion to chat and shortcuts, we have showcased the power and potential of Generative AI. However, for AI to truly augment human capabilities and enhance the overall user experience, it needs to become a bit more invisible and hidden under the hood.
Shortcuts and similar approaches will be the next phase of exploration for Machine Learning Engineers and UX Designers in the realm of text generation. It's time to get cracking!