GitLab UX Team, 13 Apr 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: First thoughts on AI UX Framework (2023-04-13)

Description

Pedro Moreira da Silva, Staff Product Designer, shares the need for a UX framework that connects all AI-driven experiences across GitLab.

Epic: https://gitlab.com/groups/gitlab-org/-/epics/10334

A

Hi everyone, my name- is Peter murder, the Silva and I'm the staff product designer working in the AI integration working group and I want to share with you um some questions, some things that have been running on through my Minds as we think about how AI will work across gitlab and right now we have a bunch of different um prototypes that are being designed and built like the Tanuki Bots that we can see here. We have the explain this code, which we can see here.

A

We have forecasting, although it's not I, don't think it's going to be using AI specifically, but it's being thought about in that process. We also have how to summarize merge requests, changes.

A

um And so it also explains this vulnerability, which is being designed here um and in this context there are so many things that could be reused by all of these different teams and I feel that that needs to be a connecting thread through the whole experience.

A

I was looking at what uh Microsoft's release with copilot and they basically came up with this very simple to understand framework um the ux framework, which is based on different altitudes, so three different altitudes immersive whole knowledge base, Focus assistive, which is in-app, focus and embedded single entity focus, and so you go from the most detailed and micro level up to the global macro level for AI interactions in the middle. You can see here the assistive with this very small mock-up in low Fidelity.

A

That shows a chat interface next to the work that you're doing, which is what we've been mostly thinking about um here at gitlab, with some of our prototypes, but overall there's a lot of work that Microsoft put into a framework that can be reused and is consistent across their product Suite um in embedding co-pilots uh having it on the side, as you work on something else uh using it for um summarization, for example, in emails, you can click here to summarize a conversation, and it would you know pop in like this, and then you can insert this summary as another email in the email thread for everyone to see.

A

So this is one approach where they basically um took all of this and created a framework for copilots, so co-pilot is the identity, is the unifying language that people can relate when they're using Microsoft products and all of the products will will have a copilot button and we'll be able to so. This is Microsoft. Word here is Powerpoint. They have the same icon.

A

um It triggers the same kind of interactions and is consistent again, all with the copilot identity, linking all AI interactions across the product Suite.

A

um Here it's Google and Google doesn't have a lot of public um information on uh their ux strategy and framework for AI. This is the only thing I could find and they don't give a specific name to the um the AI capabilities like Microsoft does with copilot. They just in this case, have something like help me write and has an icon with a pencil and a a star. You can also see here um something like this to generate images in Google Slides.

A

It uses a similar kind of iconography, and all of that is just very integrated into the actions that users are taking, so it doesn't have the AI as the at the Forefront it's more in the background and it's more integrated into what users want to do. So this is just. This is also another example sap where they have the sap digital assistant. In this case it's very minimal.

A

They only have this example with a um like a sidebar, an overlaying dialogue that you can interact and chat with a digital assistant, um but anyway, just wanted to give you an overview of how these examples show different ways of of showing how AI can be interacted with across a platform, and this is what I want us to think, and something I like to work on for gitlab is a ux framework for all of our AI capabilities and putting that Under, One Roof and one kind of experience that could solve a lot of the problems and address a lot of the questions and challenges that we've been getting from different teams working on the different, um the different prototypes, so that it is a consistent, AI experience.

A

um This is, you can think of it as the materialization of what um Katie McCoy started with these guidelines for AI human interaction.

A

You have these guidelines, which almost act as principles across everything that we do in this space between Ai and human interaction and aux framework for AI in gitlab would be the materialization how we translate this and put it into practice. So this is what I'm thinking this is what I hope to start working on and thanks for watching and let me know if you have any questions.