Parloa Inc. has joined OpenAI’s program for testing unreleased models, expanding a customer-service AI partnership that aims to improve agents and connect Parloa with more enterprise buyers.
Parloa announced the expansion Nov. 11. It said its participation gives the company a chance to evaluate OpenAI models before they become publicly available for possible use in customer service.
Engineers from the two companies expect to meet in person several times, focusing on agent response speed, reliability and operating costs. The partners said experience from customer interactions would help guide the engineering work.
Marc Manara, OpenAI’s Global Head of Startups, said the goal is to make service interactions “faster, more reliable, and more useful.” The announcement did not report measured performance results.
The companies also plan to approach enterprises interested in using AI for customer service, with North America a particular focus. Parloa’s announcement said the partnership is intended to support its international growth.
The partners expect Parloa’s AI Agent Management Platform to become available through the OpenAI Marketplace, where enterprise customers could discover it. The announcement gave no listing date or commercial terms. For Parloa, the stated benefits include access to pre-release models and a planned route to greater visibility with enterprise buyers.
The agreement extends an existing product relationship. In a May 7 profile, OpenAI said Parloa applies its models at multiple stages of enterprise voice-service systems, including testing and live operation.
Parloa also said its engineering team uses Codex in software development and that it plans to broaden the use of OpenAI tools across the company. The announcement did not identify which models Parloa would evaluate, set a testing schedule or name customer deployments or performance targets.
