Two former Google executives have successfully launched an $11.3 million fund aimed at supporting early-stage AI startups. Their conviction is rooted in the belief that the period of trial and error in enterprise AI is coming to a close, with customers increasingly opting to invest in products that demonstrate tangible value.
BAG Ventures, co-founded by Bonita Stewart, previously a vice president at Google, and Jackson Georges Jr., a former partner at CapitalG, has completed the fund after approximately two years of initial investments. The firm has already chosen to invest in ten innovative companies, including SXD, a software developer, BizTrip, an AI-powered travel assistant, and Nomadic, which specializes in agentic reasoning platforms.
The fund seeks to empower startups focusing on various sectors such as AI infrastructure, computational technologies, edge AI, security, governance, and vertical SaaS. Investment amounts vary between $100,000 and $500,000, with plans to allocate the remaining capital over the forthcoming two years.
Stewart dedicated 17 years to her role at Google, spending nearly a decade as a vice president while also holding a board position at Gradient Ventures, Google's early-stage AI fund. She actively participates as a limited partner in the Female Founders Fund and the Operator Collective. Alongside Georges, she co-led the angel network BAG Collective, which boasts over 450 members.
Georges, who previously worked at GE Healthcare and Google, where he first crossed paths with Stewart, later became a partner at CapitalG, Alphabet's growth investment division. He and Stewart were participants in the inaugural cohort of the Black Venture Institute at UC Berkeley.
The duo emphasizes their unique advantage: access. They established BAG Ventures to bridge the gap they perceive between founders and operational leaders in the AI space. "Founders need direct access to the enterprises they aim to engage with, and many high-level operators are eager to assist these early-stage entrepreneurs but lack the means to do so," Georges stated. This distinct approach not only provides financial backing but also facilitates meaningful introductions to potential customers and offers practical go-to-market guidance. BAG Ventures counts Google, alongside professionals from Nvidia, Amazon, and Snowflake, among its more than 150 limited partners drawn from a diverse array of companies.
Georges pointed out a significant evolution in enterprise purchasing behaviors concerning AI solutions. The experimental phase is being replaced as organizations focus intently on unit economics. "Enterprises are prioritizing results-driven solutions rather than investing in generic, open-ended chatbots," he remarked, highlighting the value that comes from solutions seamlessly integrated into existing operational frameworks, like automating code evaluations and analyzing legal documents.
As enterprises shift away from per-user subscriptions for software tools, Georges anticipates a future where purchases are based on completed tasks and outcomes from sophisticated workflows. To that end, BAG Ventures is keen to back core technical teams that have a history of collaboration, possess a minimum viable product, and ideally include at least one partner, alongside an indication of a rapid monetization strategy.
Additionally, the firm seeks entrepreneurs whose products can deeply penetrate enterprise processes and secure proprietary data that isn't readily available for public access. As leading AI labs begin to introduce more products independently, having a technically sound offering may not suffice for long-term success. “If a startup simply layers a thin interface over a frontier model API, they may soon face extinction,” he warned. Thus, they focus on companies that establish control over the intent layer and foster customer loyalty to withstand the wave of new model releases.
BAG Ventures is also considering initiatives in highly regulated sectors, which often require specialized compliance with data privacy regulations. “This involves securing internal data streams, creating acceptable-use policies, and implementing ongoing automated security assessments,” Georges explained, showcasing a successful implementation in their portfolio through a company called Defendremate.
In addition, Georges pointed out the increasing necessity for Identity and Access Management solutions tailored for non-human entities, including AI agents. "Startups capable of designing advanced 'zero trust' architectures and orchestration frameworks specifically for agentic systems are positioned to address a significant and lucrative market need," he concluded.
