FRUIT Introduces PAVE, an Approach to Bringing Human Judgment Into AI Systems

fruit-introduces-pave,-an-approach-to-bringing-human-judgment-into-ai-systems
FRUIT Introduces PAVE, an Approach to Bringing Human Judgment Into AI Systems

As AI systems take on more work, PAVE introduces a way for human judgment, authority, and attribution to be represented within agentic systems.

FRUIT introduces PAVE, a protocol in development designed to make human judgment accessible to AI systems through machine-readable Judgment Lenses, enabling informed decisions while maintaining permissions, accountability, and attribution.

Miami, FL, Oct. 13, 2026 (GLOBE NEWSWIRE) — FRUIT Holdings, Inc. today introduced PAVE, a protocol that gives AI and agentic systems a way to call on human judgment and authority when needed.

FRUIT Introduces PAVE, an Approach to Bringing Human Judgment Into AI Systems

FRUIT Holdings, Inc. introduces PAVE, a protocol designed to make human judgment accessible to AI and agentic systems while preserving authority, attribution, and provenance.

Cloudflare reported in September that daily requests from AI agents on its network had grown by more than 1,700% over the prior year, creating a new class of software acting on behalf of people. As agents take on more active roles across the internet, questions of human judgment, oversight and authority are becoming practical concerns for the systems being built and the public. 

PAVE is designed so software can request and apply human judgment at the moment it is needed in a workflow. Rather than treating human involvement only as a person waiting in the loop to approve an action, PAVE explores how judgment itself can become callable while preserving whose judgment is being used, the authority under which it is applied and its connection to the person it came from.

“AI is getting very good at doing things,” said Angela Benton, founder of FRUIT. “But removing a person from a workflow shouldn’t mean removing them from the value, authority or judgment they contributed to it. We’re interested in how agentic systems can account for human judgment and authority when they matter, and how to do that without separating those contributions from the people they came from.”

FRUIT calls the broader concept Callable Human Judgment: human-derived judgment represented so a system can request its application when needed. PAVE is the company’s first system built around that idea.

Human oversight, safeguards and intervention are already part of the public, technical and policy conversation around AI. FRUIT’s work focuses on what comes next: creating practical ways for those ideas to be represented within systems.

“The public conversation around AI tells us that people are concerned about where they fit as these systems become more capable,” Benton said. “Technology is going to keep moving forward, and as it does we have an opportunity to build systems that account for people appropriately and fairly; including human judgment, authority and attribution as part of the architecture.”

PAVE is developed for environments where organizations, systems, or governing policies determine that human judgment or authority should inform a decision. Potential applications span enterprise operations, AI evaluation, and agentic workflows, including areas such as financial services, healthcare, and the public sector. PAVE is designed to interoperate with existing systems rather than replace their decision-making or operational infrastructure.

Learn more about PAVE at byfruit.io/pave and explore the public specification repository.

About FRUIT Holdings, Inc

FRUIT Holdings, Inc. is an origination company that develops and commercializes original technologies and infrastructure where new systems, approaches, or business models are needed. Its current work includes PAVE, a protocol designed to make human judgment accessible to AI and agentic systems while preserving authority, attribution, and provenance.

Press Inquiries

FORME Communications
hello [at] formecomms.com

A video accompanying this announcement is available here: https://youtube.com/watch?v=3zclG362kCw