OpenAI's launch of Presence represents a direct challenge to the incumbent CX platform vendors who have spent years building agent capabilities into their ecosystems. The platform consolidates voice and chat workflows under a single AI agent interface, positioning itself as an alternative to the agent modules within Zendesk, Salesforce Service Cloud, and similar suites rather than as an add-on to them. This matters because it signals OpenAI's intent to compete for the enterprise CX dollar not through partnerships or integrations, but through displacement—offering teams a purpose-built alternative that bypasses the need for expensive, feature-bloated CX platforms if their primary use case is agent automation. The critical question for support leaders already embedded in these ecosystems is whether Presence's focused approach to voice and chat will prove sufficiently compelling to justify the operational friction of ripping out existing workflows, or whether the switching costs and integration complexity will keep most teams locked into their current vendors despite superior AI capabilities.
The timing of this launch reflects the broader market reality that AI agents have moved from experimental to operational, and that CX teams are now evaluating them on practical metrics: deployment speed, accuracy, and cost-per-interaction rather than novelty. Presence enters a market where Gartner has already cautioned CX leaders against treating AI agents as employees, meaning the conversation has shifted from "can we build this?" to "how do we govern this responsibly?" For teams currently running Agentforce or similar vendor-native solutions, Presence forces a strategic choice: does the marginal improvement in AI quality justify the operational overhead of managing a separate platform, or does the integrated approach of your existing vendor—however imperfect—remain the path of least resistance? The answer likely depends on whether your organisation has already built the governance frameworks and monitoring practices that self-learning AI systems require, or whether you're still in the early stages of agent deployment where standardisation matters more than optimisation.
OpenAI Launches Presence, an Enterprise AI Agent Platform for Voice and Chat Workflows MLQ.ai