Resect AI Secures $25 Million Funding to Combat AI Hallucinations with Real-Time Detection Technology

Resect AI Emerges with Real-Time Solution to Prevent AI Hallucinations

Resect AI launched Thursday from stealth mode with $25 million funding to commercialize technology that addresses one of artificial intelligence's most persistent challenges: AI hallucinations. The Seattle-based startup takes on AI hallucinations through patented in-stream technology that operates fundamentally differently from existing solutions1

1

2

While traditional monitoring tools evaluate generated text after fabrications occur, Resect AI's platform works in real time, looking deep inside large language models to observe, detect, interpret, and modify model behavior before confabulations happen. This approach creates an accountability layer for enterprise AI by intervening directly within the model's internal decision-making process rather than running post-hoc checks1

1

Source: SiliconANGLE

Understanding the Technology Behind Preventing AI Hallucinations in Real Time

The company's journey began with building its own AI models capable of producing high factuality scores. Through controlling data, developing training processes, and implementing behavior and reinforcement training, Resect AI discovered these techniques could be applied to existing open models including DeepSeek, Qwen and Llama2

2

Chief AI Officer Tim Walton explained the company's research focus: "We've spent an extensive amount of time and resources researching how models think, and what causes them to choose the answers that they do. Through this process, we've developed technology that observes exactly when and how models fail, and surgically fixes them"2

2

Currently, Resect AI offers two models on HuggingFace: a 0.6-billion-parameter Veritas fact checker and an 8-billion-parameter model. Both utilize the Qwen3 architecture without underlying chain-of-thought layers. The 0.6B model achieved an average score of 72.3% on LLM-AggreFact evaluations, representing a 7.4% improvement over standard Qwen32

2

Building Enterprise Trust Through Audit and Modify Model Behavior Capabilities

Resect AI's internal visibility tools form the foundation of the NeuroWave Product Suite, described as a polygraph for neural networks. The platform stops fabrications at the source while simultaneously generating audit trails for enterprise compliance and due diligence1

1

CEO Kevin Owens emphasized the urgency: "AI has prematurely been put in a position of trust. Adding labels such as 'use at your own risk' flies in the face of proper governance or compliance. We are building the next large enterprise AI company to bring transparency and accountability to AI for industries such as publishing, finance, healthcare, research, and education where factual accuracy is absolutely critical".

Source: GeekWire

Related Stories

Scaled Cognition raises $100M from Khosla to eliminate AI hallucinations with new architecture

Probably AI Raises $9M to Catch Hallucinations Using Smaller, Cheaper Models

Vectara Unveils Guardian Agents to Combat AI Hallucinations in Enterprise Applications

Moondream Raises $4.5M for Compact Vision Language Model, Challenging AI Industry Norms

Strategic Location and Talent Acquisition Plans

Resect AI established its headquarters in Washougal, Washington, a city of approximately 18,000 residents located 175 miles south of Seattle along the Columbia River. The company employs four people at its Main Street office out of a 30-person workforce distributed across the Seattle area, California, New York, and Texas1

1

Kevin Owens noted the location strategy has already enabled talent acquisition from both the greater Seattle and Portland markets, with PhDs joining from the regional talent pool. The startup plans to open an additional office in the Seattle area for engineering and business operations1

1

Funding Allocation and Go-to-Market Expansion

The $25 million funding from private equity investors will accelerate research and development, expand go-to-market initiatives, and fuel talent acquisition. Resect AI aims to grow its headcount from 30 to 50 employees by the end of 20261

1

2

The leadership team includes Kevin Owens as CEO, Tim Walton as Chief Artificial Intelligence Officer, Tyler Gerber as Chief Operating Officer, and Tommy Lofgren as Chief Product and Marketing Officer. The company plans to release open-source tools through a GitHub repository to support enterprise adoption2

2