Scaled Cognition just raised $100M to kill the one thing killing your AI deployments

Your AI works 70% of the time.

That 30% failure rate is why your agents stay in pilot.

Scaled Cognition just closed a $100M Series A led by Khosla Ventures. The pitch: a model that will not give you a wrong answer.

Not "reduces hallucinations." Not "improves accuracy." Eliminates them.

The model is called APT — Agentic Pretrained Transformer. It's already in production with Fortune 500 enterprises across financial services, healthcare, telecom, and insurance. Genesys — serving 8,000 organizations in 100+ countries — is both a customer and an investor.

Here's what matters: the founders already sold an agentic AI company to Microsoft. They know what enterprises actually need, not what demos suggest.

The target is the $600 billion business process outsourcing market. Customer service, IT support, HR, finance. Enterprises are ready to insource these functions — but only if the AI doesn't hallucinate a wrong bank balance or botch an insurance claim.

Over the next 12 months, companies using Scaled Cognition's models are on track to automate more than one billion customer service interactions.

"When a system makes mistakes 30% of the time, complex issues go unresolved and customers don't come back," said CEO Dan Roth. Their models resolve most issues, saving hundreds of millions in operational costs.

The real signal: enterprises aren't stalling on AI because of capability. They're stalling because of trust. Scaled Cognition is selling the architecture that turns capability into reliability.

If your AI deployment is stuck in pilot, this is why. And this is who fixes it.

Audit your AI vendor's hallucination rate today. If they can't give you a number, they don't have one.

SOURCE: https://thenextweb.com/news/scaled-cognition-100m-khosla-reliable-ai
VERIFIED: GlobeNewswire (company announcement), The Next Web, The Wall Street Journal
SIGNAL: Reliability is the #1 blocker to enterprise AI adoption. Scaled Cognition's $100M raise at $750M valuation proves investors are betting that architecture — not more compute — solves the hallucination problem. Your AI vendor's failure rate is now a competitive liability.