Seed Round Signals a New Era in AI-Driven Discovery
Seltz, a New York–based startup focused on rebuilding search for AI agents, disclosed a $12.5 million seed round this week. The round was led by Speedinvest and B Capital, with participation from the Italian Founders Fund, United Ventures and Future Back Ventures, the Bain & Company venture arm. The investment underscores how investors see a growing need for data surfaces that AI agents can cite and reason over, not just quick snippets.
Why Seltz Says Traditional Search Falls Short for AI
Founder and CEO Antonio Mallia argues that conventional search engines were designed for humans typing brief queries and scanning a ranked list of links. AI agents, by contrast, emit long, precise queries and often run dozens of questions in parallel. They require machine-ready information that a model can cite, not a human-friendly snippet meant to drive clicks.
Mallia notes that the relevant data for AI workflows lives deeper in pages—tables, structured content and images—rather than in the top result snippet. He says the traditional engine's architecture makes it hard for an agent to fetch the kind of evidence modern AI systems demand. The exclusive: seltz, startup rebuilding narrative around this shift is gaining traction as agents become more capable in research and decision tasks.
What the Funding Tells Us About the AI Search Race
The seed capital positions Seltz as one of several players attempting to redefine search for AI-powered assistants. In a market where firms race to build AI-native discovery layers, the emphasis is shifting from keyword matching to data-rich surfaces that can be cited by large language models. Seltz aims to surface content from internal data sources and public pages in a format that AI models can anchor to, including structured data and media assets.
“The old search methods were architected for humans,” Mallia said. “The information AI agents need sits in the body of the page and in representations that a model can reason over, not in a catchy snippet.”
Leadership and Roadmap
Mallia, who holds a PhD in information retrieval from New York University, previously worked as an applied scientist on Amazon's AI initiatives and as a research scientist at Pinecone, the vector-database specialist. He frames Seltz as a response to a decades-old design problem: retool search for the needs of AI workflows rather than human browsing patterns.
The company says the seed funds will accelerate product development, expand data partnerships and broaden support for AI agents that operate across disparate domains—from consumer products to financial services. The plan includes enabling agents to access richer, machine-readable content and to cite sources confidently in real time.
Investor Reactions
Speedinvest and B Capital, both active in cross-border tech rounds, described the moment as a meaningful step in the AI search ecosystem. A partner at Speedinvest said the field needs new data surfaces and robust curation to satisfy AI-agent workflows, especially in regulated or high-stakes sectors. A B Capital executive noted that successful AI agents rely on trustworthy, traceable content, which traditional search often struggles to deliver at scale.
Other participating funds, including Italian Founders Fund, United Ventures and Future Back Ventures, cited the European and US technology feedback loop as a key driver for backing a founder who blends research depth with operator experience. The round adds to a growing list of seed rounds aimed at building infrastructure for AI agents and conversational systems.
Implications for Personal Finance and Everyday Consumers
For consumers and personal finance enthusiasts, the shift toward AI-native search could shorten the path to accurate, up-to-date product comparisons and financial education. AI agents that can pull from official product docs, fee schedules and risk disclosures—while clearly citing sources—could help users make more informed decisions about credit, investments and insurance. In this sense, the exclusive: seltz, startup rebuilding trend resonates with readers seeking transparent, machine-verified information at the point of decision.
Already, dozens of fintechs and consumer fintech sites rely on rich data feeds for AI-assisted advice. A more robust, machine-readable search layer could reduce the time it takes to surface credible information and could improve the trustworthiness of automated financial recommendations over time.
What Comes Next
Seltz plans to deploy the seed capital over the next 18 months to advance core capabilities, expand its content partnerships and pilot with select AI agents that require high-quality, citeable data. The team aims to launch a beta for a limited set of AI agents in the second half of 2026, with broader deployment anticipated as data surfaces mature.
In a fast-evolving market, the company argues that the next generation of search will be less about listing results and more about delivering verifiable evidence that AI can explain and trust. The focus will be on building robust provenance, improving data normalization across sources and enabling agents to reason with complex content such as tables and charts.
Key Data Points
- Seed round size: 12.5 million USD
- Lead investors: SPEEDINVEST, B CAPITAL
- Other participants: Italian Founders Fund, United Ventures, Future Back Ventures
- Founder/CEO: Antonio Mallia, PhD in information retrieval from NYU; former Amazon AGI and Pinecone scientist
- Purpose: rebuild search to support AI agents and their research workflows
- Timeline: beta rollout planned for second half of 2026
Discussion