Two Google alumni have successfully raised a fund of $11.3 million to support early-stage AI startups. They believe that the "experimental" phase of these companies is coming to an end, and that customers will increasingly be willing to pay for products that can truly prove their value.

BAG Ventures was founded by former Google Vice President Bontia Stewart and former CapitalG partner Jackson Georges Jr. After about two years of preparation, the fund officially completed its fundraising process. To date, the company has invested in 10 enterprises, including software companies SXD and AI, travel agency BizTrip, as well as platforms agentic and reasoning.

The startup companies invested in by this fund cover areas including AI infrastructure, computing power, physical AI and edge AI, security, governance, as well as vertical SaaS. The size of each investment ranges from $100,000 to $500,000, and the team hopes to invest the remaining funds within the next two years.

Stewart has worked at Google for 17 years, during which nearly 10 years she served as a vice president. During that time, she also served as a member of the board of directors for Gradient Ventures, which is an early AI fund of Google. She is also a limited partner in Female Founders Fund and Operator Collective. Together with Georges, she led the angel investment consortium BAG Collective, which has over 450 members.

Georges used to work at GE Healthcare and Google, and it was during his time at Google that he met Stewart. Later on, he became a partner in Alphabet's Growth Fund CapitalG. Both he and Stewart were among the first batch of students at Berkeley Black Venture Institute.

The two individuals stated that their advantage lies in their “contact channels.” Georges said that they founded BAG Ventures in order to “bridge the growing AI gap between founders and operators.”

"Founders need to get inside the organizations they want to sell to, and we know many senior operators who are willing to support early-stage founders, but don't know how to do it," he said. Therefore, "we don't just provide funding for founders; we also introduce them directly to potential customers and offer [go-to-market] advice through our personal involvement," he added. The limited partners of this fund include Google, as well as operators from NVIDIA, Amazon, and Snowflake; there are more than 150 limited partners in total, spread across numerous companies.

Georges said, "Many funds have operator networks. We hope to be the one that is truly useful."

The investment logic of Georges is based on his judgment of the changes in the way enterprises purchase AI. He said that the "experimental sandbox" phase is coming to an end.

"Companies are now paying great attention to the economic efficiency of their units," he pointed out. "They no longer just pay for open-source chatbots; they are willing to pay for solutions that provide certainty. The real value comes from those solutions that can be deeply integrated into traditional work processes and actually get the work done." For example, automated code review and the parsing of legal documents.

Georges is preparing for a world where businesses no longer purchase SaaS tools per user seat. "We will be purchasing completed tasks and results driven by multi-agent workflows," he said.

For this reason, BAG Ventures hopes to invest in companies that have had core technical teams working together, already have a minimum viable product, have at least one partner, and "have a very clear path to monetization within 24 hours."

He also hopes to bet on products that penetrate deep into corporate workflows and can access proprietary data that cannot be scraped. As cutting-edge AI laboratories launch more products of their own, Georges believes that merely creating a technically viable startup product is not enough in the long run; “If a startup is just a thin layer of packaging around the cutting-edge model API, they will quickly be eliminated.” This is also why they are looking for teams that penetrate deep into corporate workflows and have access to proprietary data that cannot be scraped. “We want companies to have an intention layer and sufficient customer lock-in capabilities to survive the next major model release.”

The fund is also focusing on startups that sell products to highly regulated industries, where data privacy requirements may necessitate a higher level of specialization. “This means protecting internal data flows, establishing acceptable safeguards, and deploying continuous automated red-team testing,” said Georges. “We have seen that this approach has worked well in our portfolio company Defendremate.”

Georges also indicates that companies will need access management tools for non-human employees, such as AI agents. "Startups that are capable of building the next layer of 'zero trust' architecture and orchestration channels for agentic systems will fill a huge and very lucrative gap," he said.