Box Inc. (NYSE:BOX) CEO Aaron Levie said AI agent adoption and deployment are being held back by challenges in testing, tuning, and optimizing agents in real-world work environments.
AI Agent Testing Faces Enterprise Bottleneck
On Monday, Levie said in a post on X that companies face challenges testing AI agents across files, customer relationship management systems, email platforms, and other tools they use in their daily operations
"A decent amount of agent adoption and deployment is bottlenecked by being able to test, tune, and optimize agents on ‘real’ work environments," he said.
He also said companies need to understand how well agents perform on evaluations, or "evals," and how they could perform after a model change or workflow upgrade.
"Every enterprise has to do this 1 by 1 right now, which is obviously slow and tedious," Levie said.
He added that companies will need dedicated resources to manage evaluations and build infrastructure for running tests and simulated environments.
"Every enterprise is going to have someone managing evals, infrastructure for building and running evals and simulated environments, and more," Levie said. "Huge space."
AI Agents Reshaped Software
Last month, Coinbase Global Inc. (NASDAQ:COIN) CEO Brian Armstrong said AI agents could make internal software platforms compete based on performance, cost, and reliability, while allowing employees to compare and switch between services.
In August, Venture capitalist Chamath Palihapitiya said AI agents and data security concerns could weaken the traditional bottom-up software buying strategy as companies sought to prevent proprietary information from leaking through unauthorized AI tools.
Earlier this year, former Tesla AI chief and OpenAI researcher Andrej Karpathy said AI agents had transformed his coding workflow, with AI systems handling most of his programming.
He also built an AI assistant called "Dobby" to manage various smart-home functions.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Image via Shutterstock
Login to comment