NOTE: This Thursday’s holiday party is the GenAI Collective's final event of the year and our milestone 50th event! Come join in on the festive fun and toast to our collective accomplishments!!

The user interface in Generative AI is rapidly evolving, moving beyond the familiar chatbot model to more integrated, context-aware, and multimodal experiences. Chat interfaces, such as ChatGPT, brought Generative AI to the mainstream. However, they often place the onus on users to draft effective prompts, interpret responses, and then produce the necessary content across the application of their choice. Many of us have dealt with the frustration of using ChatGPT to produce a thoughtful answer, and then being forced to spend significant time transforming the content into a consumable medium (adding business context, moving to another application, validating accuracy)​​. As generative technologies become more ubiquitous and advanced, there's a growing need for startups to mature past the co-pilot phase and complete the end-to-end workflow.

We have already witnessed the vast potential of Generative AI across virtually all functional areas of business, ranging from developing creative marketing materials to back-end infrastructure or data pipelines​​. Generative platforms of the future will need to integrate with business applications and maintain greater contextual awareness. For instance, in February 2023, Adept AI raised a $350M Series B with the promise of building a machine learning model that can transform the way knowledge workers interact with computers. The company’s vision moves past the traditional chat interface or co-pilot paradigm to offer end-to-end actions across the landscape of enterprise applications (Office 365, Salesforce, JIRA). Additionally, most developers are already using the toolset offered by GitHub Copilot because of the platform’s ability to integrate seamlessly into user workflows, providing suggestions and accelerations without disrupting the natural work process​​. In fact, in mid-2023, data from GitHub showed 41% of code was AI generated (CNET). This next frontier of Generative AI requires the platform to have a deep understanding of the user's current context, history, and workflow choices​​.

The rise of multimodal experiences, including voice and gesture interfaces, is making generative technologies more ubiquitous in our lives. For example, Humane's vision of a screenless future where interaction occurs through natural language points to a significant shift in how we will interact with the physical world​​. Instead of relying on traditional apps, Humane believes natural language alone will be the primary interface for interacting with software. Generative AI should enhance user creativity and control, especially in the design and creative domains. Tools like Midjourney, DALL-E, and Adobe Firefly democratize creative expression, but it's equally important to provide users with the tools to edit and manipulate outputs to suit their needs​​.

A significant barrier in leveraging LLMs and AI models effectively is mastering prompt engineering. Education and awareness on how to engage with these models for specialized tasks are essential​​. Additionally, the utility of an AI model hinges on its ability to provide meaningful and accurate results. Training models on datasets that closely represent real-world workflows (rather than just data) and covering a wide variety of use cases is crucial​​. It's also important to build mechanisms that remind users of the limitations of AI models, especially when the results could influence critical decisions​​. Some early examples of practical implementations include tools like DawnGPT and Diana developed by the management consulting firm the Nous Group. DawnGPT acts as a natural language interface to a SQL database, translating user queries into SQL and presenting results in various formats. Diana is a logic tree generation tool aiding in structured problem-solving, highlighting the shift towards task-specific, user-friendly AI tools​​.

The evolution of interfaces in Generative AI is marked by a move towards more integrated, contextual, and multimodal experiences. This shift promises to make Generative AI more accessible, intuitive, and effective in solving end-to-end workflows and enhancing creativity. As Generative AI continues to advance, the focus will increasingly be on designing interfaces that are not only technically capable but also deeply attuned to user needs and contexts.

As we continue to see this ecosystem evolve, we would love to hear your perspective on how you see the Generative AI user experience developing. If you have any ideas or insights you would like to share, hit up Eric on the community Slack or reach out via email at [email protected] for a feature in our next newsletter!

Last week, we were excited for some additional firsts! Thanks to our co-hosts Coco Chen and Matt Huang for making it all possible, and to our incredible attendees for participating in what became a lively and intellectually stimulating evening! 🥳

About Eric Fett

Eric joined The GenAI Collective in early September to lead the development of the newsletter. He is currently an investor at NGP Capital where he focuses on Series A/B investments across enterprise AI, cybersecurity, and industrial technology. He’s passionate about working with early-stage visionaries on their quest to create a better future. When not working, you can find him on a soccer field or at a sushi bar! 🍣