Seltz, which is building a web search engine that can be used by AI agents, raised a $12.5M seed led by Speedinvest and B Capital
The rise of AI has rekindled the long-dormant search wars. Chatbots and AI agents need to surface timely, relevant information about news and all kinds of products and services.
Context & Ripple Effects
Seltz enters a small but increasingly well-funded category of search products designed specifically for AI agents. Related coverage shows Exa raising a much larger round at a sharply higher valuation only weeks earlier, signaling that investors see agent-oriented retrieval as a distinct infrastructure layer rather than a chatbot feature alone.
The wider coverage also includes Browser Use, which translates web interfaces into forms agents can interpret, and ZyG, which coordinates agents for marketing and SEO work. Together, those companies frame Seltz’s product as part of the stack that lets agents find, understand, and act on web information.
First-order effects
- Seltz gains $12.5M of seed capital to build its agent-facing web-search product, with Speedinvest and B Capital backing its attempt to compete for developers and agent builders.
- AI-agent teams have another potential retrieval provider for obtaining current web information, rather than relying solely on general-purpose search or model-internal knowledge.
Second-order effects
- Exa and other agent-search providers face stronger pressure to differentiate on the relevance, freshness, and usability of results for automated workflows, not just conventional search quality.
- Tools that help agents interpret websites or orchestrate marketing tasks become more complementary to dedicated search: better retrieval can improve the inputs those products use to navigate sites and make decisions.
Third-order effects
- If agent use expands, web search may be reorganized around machine-consumable retrieval APIs and agent workflows, alongside the consumer search interface rather than beneath it.
- That shift would make access to timely web data and reliable interpretation of web content a more consequential control point; the eventual winners will depend on whether specialized providers can sustain an advantage over broader AI and search platforms.
The trend: This is one data point in the emergence of an agent-native web infrastructure stack, spanning retrieval, web understanding, and task orchestration.