In early 2023, David Petrou, a distinguished engineer and founding member of both Google Goggles and Google Glass, made a surprising move. After more than 17 years at the company, he departed to launch his own startup.
“I was seeing how fast technology was changing, and I felt there are certain ideas that are best explored in the context of a startup,” Petrou told TechCrunch.
His ultimate idea was to build Continua, a consumer-facing company that uses AI agents to enhance collaboration and interaction in group chats on SMS, iMessage, and Discord.
“The simple way to think about it is that we’re bringing the power of LLMs to group chats,” Petrou said.
Continua announced on Tuesday that it has raised an $8 million seed round. The funding was led by GV, with additional participation from Bessemer Ventures Partners and a group of angel investors.
When developing Continua, Petrou noticed that people would interact with ChatGPT or another LLM, and then would copy and paste what they learned into their group conversation.
“If you and I are planning a trip, or if we’re trying to figure out what to have for dinner, or what movie to watch, all of these things can be facilitated by an LLM participating in a group chat with us,” Petrou said.
Techcrunch event
San Francisco
|
October 27-29, 2025
Continua claims its AI agents can reduce group chat chaos by joining conversations and offering helpful information only when it’s needed.
As the group discusses projects and common plans, Continua can automatically set reminders, launch polls, add calendar invitations, or generate Google documents with checklists and to-do lists.
If a user forgets details from the group chat, such as the meeting time or location, they can simply direct message Continua to privately ask for the information.
At first glance, Continua may seem like a straightforward application of AI, but Petrou says that getting LLMs to participate in a conversation with multiple humans is a rather complex technical problem. Since most AI models are designed for conversations between a single person and a single assistant, Continua had to fine-tune LLMs to understand the d …