New offering gives companies a way to measure how people actually relate to their conversational and agentic AI, beyond accuracy and technical performance

BELLEVUE, Wash.: Auros, the human network behind AI, today announced AI Relationship Quality (ARQ), a new benchmark that diagnoses how well an AI experience is working for the people using it. Delivered by Auros’ Professional Services team, ARQ extends the traditional AI eval model, which tests accuracy, efficiency, and safety, to also test human relationship elements like trust and affinity.

Auros ARQ closes the gap between technical performance measures and the business results that AI-powered experiences are intended to deliver. The ARQ score is a reliable, quantified way to know whether an AI experience actually works for customers: whether they understand it, trust it, feel in control, get the outcome they need, and want to keep using it.

“You can have the most technically capable AI on the market and still lose the customer,” said Jason Giles, Vice President of Customer Intelligence, Auros. “Because people aren’t grading your model or agent, they’re deciding whether to trust it. ARQ exists because that decision shouldn't be a guess. It’s a difference between assuming your AI is working and knowing it is.”

How ARQ works
ARQ is an expert-led benchmark and assessment conducted by Auros’ Professional Services team using a rigorous, repeatable methodology designed for consistent measurement over time. Each engagement is tailored to a company’s specific AI experience, target audience, and business goals, with participants recruited to match.

The benchmark produces one overall ARQ score, plus individual scores across the five elements that drive AI adoption and satisfaction:

• Understanding - Do human users feel they understand the LLM’s thinking, intent, and boundaries?
• Trust - Do people give the appropriate level of trust – not too much, but also not too little -- to an AI experience?
• Control - Can people steer the AI, correct it, and recover when it goes wrong?
• Outcome - Does a user feel the AI-human partnership delivered better results than the user could alone?
• Affinity - Does the AI’s personality, tone, warmth, and emotional response make people want to come back?

Because the benchmark is repeatable, it can be used to measure progress over time and compare against competitors. The scores are paired with video feedback from participants, giving teams the context behind the numbers: what’s working, what needs improvement, and what to prioritize next. The result is a comprehensive view of the human-AI relationship, including findings and recommendations.

ARQ is designed to complement AI evals, not replace them-where evals measure system performance, ARQ measures the human relationship with the system, giving teams a full picture of whether an AI experience is actually effective.

Availability
AI Relationship Quality is available now through Auros Professional Services. To learn more or assess your own AI experience, visit aurosglobal.com.

About Auros
Auros, headquartered in Bellevue, Washington, provides the customer insights and human intelligence that product, marketing, and AI research teams need to succeed in the AI era. Backed by a network of 7.6 million verified people and two decades of recruiting and verification infrastructure, Auros is also the company behind UserTesting and User Interviews. Learn more at aurosglobal.com.

Fonte: Business Wire