← Accueil actu | Salle du marché
2026-09-10 · Source：SiliconANGLE
Seattle startup Resect AI raises $25 million in early funding to build an accountability layer that captures and reduces enterprise AI hallucinations at runtime.
Resect AI, a Seattle-based artificial intelligence research startup, announced Thursday that it has raised $25 million in early funding, marking its formal entry into the enterprise AI market. The young company is betting that one of the technology's most persistent failure modes can be managed through a dedicated accountability layer wrapped around live systems, rather than left as an unavoidable cost of deployment.
The failure mode in question is the hallucination, also known as a confabulation: an instance in which an AI model replies with a false or fabricated response and presents it with high confidence. That combination of error and assurance is precisely what makes hallucinations so corrosive in commercial settings. An output that reads as authoritative is difficult for non-specialist users to challenge, meaning fabricated facts can travel into reports, decisions and customer-facing workflows before anyone thinks to verify them.
Resect's response is an accountability layer for enterprise AI designed to capture and reduce hallucinations at runtime. The runtime element is the differentiator worth watching. Treating fabricated output as a live operational event — something to be caught in the moment a model produces it — places error handling inside the inference path rather than in downstream audits or manual review. For enterprises moving generative systems from pilots into production, that positioning speaks directly to the governance and trust gaps that most often stall deployments.
The $25 million in early funding gives the Seattle startup runway to build out that layer and test it against real enterprise workloads. The launch lands amid persistent enterprise skepticism about model reliability, which remains one of the most frequently cited barriers to scaling AI beyond pilots. Whether runtime accountability can be delivered without degrading performance is the question Resect will now have to answer in production.
Mots-clés: Resect AI · Hallucinations · Enterprise AI · AI Reliability
萬安算交所 AIXX · AIXX · Bourse compute vérifiée · Accéder à la salle