As a competitive strategy in both video conferencing and AI chatbot markets, Meet Anyone leverages Gemini Live to bring AI personas into Meet calls. In addition to Workspace enterprise usage, there are immense opportunities for new consumer adoption in education and social media spaces. At the heart of this proposal is that people are interested in meeting with those who are not usually accessible. By providing a portal to meet these personas, Google Meet can establish itself as the premiere platform to meet anyone, and garner engaging and widespread usage.
Ideation & Design Thinking In response to my Meet leadership's request for innovative and differentiating features to standout from competitors like Zoom and Microsoft Teams, I came up with this idea to make Google Meet a portal to meet anyone virtually.
Creative Direction I identified various use cases and entrypoints, AI chatbot features, user-AI interactions, considered how to adapt Gemini Live voice capabilities and Gemini Sidekick panel into Google Meet call experience, and wrote scripts for the protoyped calls.
UX/UI Design As the sole designer on this project, I executed all my concepts into high fidelity prototypes and animations, leveraging external AI facial and lip synching tools.
Product Management I created a slide deck and pitched my concept to Meet leadership, as well as the greater Google community, resulting in my current partnership with Google DeepMind team to harness their LLM technology with my creative vision for this project. I also devised a new consumer revenue model (similar to YouTube advertising) for Meet Anyone to monetize beyond the Workspace enterprise revenue model.
As of 2025, the term "AI agent" can refer to many chatbot types and functions, as shown below. This Meet Anyone proposal focuses on the AI Persona "consultants".
Below are the various types of AI chatbot personas that I am proposing users can invite to and interact with on a Google Meet call. There are further personas I have considered, but am not focusing on yet.
Meet Anyone will leverage existing Google Meet capabilities, fully accessing presentation views, dynamic layouts grid, and interactive Gemini AI side panel. Meet Anyone calls may not always utilize presentations, but the model and controller will always be present. Below is the proposed UI architecture for a Google Meet Anyone call.
Model (AI): Multiple Gemini instances can simultaneously be in the same meetings and interact with each other, and are capable of presenting differing opinions and viewpoints based on their personas.
View: Users can present, or AI models can present to users (even if they are not intending to present, relevant imagery appears to user for reference).
Controller: Users are given suggestive chips for talking points. They can edit AI model settings, or invite more AI bots. They can also enter prompts in the input field. AI models can provide links, artifacts, and resources here.
In the design prototype below, generic expert AI bots are invited into a Meet call, to help my hypothetical business partner and I prepare for an investment funding pitch. Here, these generic expert bots roleplay specific personas, ask questions, and provide feedback to me and my partner. They are able to share artifacts and reference links in the Gemini side panel.
In the design prototype below, I (hypothetically a student, researcher, or just a user interested in this subject matter) am searching for the Italian Renaissance painter Raffaello Sanzio on Google Search. The results page displays the usual wikipedia and other links about this historical figure. In addition, there is a "Meet Rafaello Sanzio" link shown, so I can interact with the historical figure via a Google Meet call.
In the design prototype below, I am meeting with professional tennis player Roger Federer via a Meet call. Similar to Twitter's verified blue check mark, public figures such as celebrities and influencers, on Google Meet would be verified and approved by Google and the said public figures for users to interact with them. In the Meet UI, there is a right side panel, with suggestive chips for users to have talking points with the public figure. Hence, public figures such as Federer can include their sponsors, charities, projects, and upcoming events as suggestive chips to promote to users.
In the design prototype below, as I am meeting with pro tennis player Roger Federer's AI bot on Google Meet, there can be natural and seamless integration of advertisements from brands he is sponsored by. This enables greater brand reach, and for the sponsored celebrity to attain more revenue from viewership. Users may not mind these ads as much since it is from a public figure they are interested in. After certain amount of seconds, there can be a "Skip Ad" button appearing as well. This revenue model, like YouTube's, benefits the ecosystem: Google Meet (since they will have a portion of the advertising revenue), the sponsored public figure, and the brand. This YouTube revenue model is proven to retain user viewership since users are willing to watch ads for content they care about. Similarly, they will be willing to view ads in order to continue interacting with public figures they are interested in.
Learn About is Google's educational LLM platform for various topics. Users are able to scroll through a list of topics (Pinterest style), and upon clicking a topic, they get brought to that topic page, where information is presented to them in progressive disclosure. Users can dive deeper into specific areas by choosing certain modules in a "Choose Your Adventure" type flow. In the examples shown below, I am proposing endless possibilities for educational conversations with AI bots via Google Meet.
Since my proposal for Google Meet Anyone has many educational use cases, it makes sense for this Meet experience to exist on other Google platforms as well, such as YouTube. There are many content creators on YouTube who provide tutorials. Currently, users comment on the video if they have questions, and would need to wait for the creator to possibly respond to them. What if the content creator had their own AI bot on their YouTube page that users can interact with and ask questions to? Check out my proposed design for this experience in the prototype video below:
I have already pitched the vision concept and design prototypes to my Google Meet leadership team, and it has received alot of positive reception and support. It was brought to the attention of the Director of PMs Dave Citron on the Gemini team, who enlisted Google's DeepMind (internal AI team) to partner with me on further developing my Google Meet Anyone feature. Stay tuned for more updates!
