See all guests

Guest Spotlight: Mayank Agarwal

Mayank Agarwal
RingCentral | Product

About Mayank Agarwal

Mayank Agarwal is Director of Product Management for AI Products at RingCentral, where he builds and measures AI agents in production. He created DART, a four-metric framework for evaluating AI agent performance, and focuses on how teams can test, measure, and ultimately trust increasingly autonomous AI features before putting them in front of customers.

Notable Quotes

"An AI agent can be accurate on the narrow task defined in the test dataset and still fail the user completely."

"When people rewrite everything that the AI gives them, you know that you have not earned the user's trust."

"It finished the job, but nobody trusted the output."

"If you keep showing someone a flash sale, then people are trained to assume that another one is always around the corner."

"You cannot A/B test a probabilistic system the way you would test a button color."

Mayank Agarwal
RingCentral | Director of Product Management, AI Products

RingCentral's DART framework: The four metrics that actually measure AI agents

Go to S1 | E15

Subscribe to the podcast

Takeaways

All Takeaways

A losing experiment is paid-for information. Groupon's flash-sale flop revealed the lever was wrong, not the goal — scarcity beat price-based urgency.

Go to S1 | E15

Acceptance rate is the trust metric. The share of output users keep without editing is the strongest available proxy for trust.

Go to S1 | E15

DART measures behavior, not opinions. Four signals read off logs and transcripts: decay, acceptance, relevance, and task completion.

Go to S1 | E15

Thumbs-up/down feedback is sparse and skewed. Unhappy users rarely rate — they just quietly stop using the product.

Go to S1 | E15

Accuracy is a comfortable lie. It grades a narrow test set and can stay high while the agent fails real users.

Go to S1 | E15
The experimentation edge podcast logo with a picture of host Ashley Stirrup