Parloa Warns Global Enterprises Are Blocking Customer Access in Landmark CX Study

New updates have been reported about Parloa.

Parloa has released what it calls the largest AI agent-led benchmark of enterprise customer experience to date, positioning the company at the center of a growing gap between automated interfaces and actual problem resolution. Using its own AI discovery agents, Parloa audited more than 10,000 enterprise websites and ran nearly 4,000 secret-shopper interactions at over 800 companies, concluding that many large organizations are structurally discouraging live customer contact despite heavy investment in digital CX.

The study, titled “State of Agentic CX in 2026,” found that 43.3% of major company websites show neither a visible support phone number nor a chat option, and that phone remains the primary channel where contact is allowed, appearing on 54.1% of sites versus 28.9% for chat. Parloa’s CMO Latane Conant said the findings confirm that enterprises still treat service as a cost center, with digital front doors optimized for deflection rather than relationship building and lifetime value expansion.

The report highlights that current chatbot deployments are largely ineffective and technologically dated, with fewer than one in eleven AI-led chat sessions resolving the customer’s stated goal and successful escalation to a human occurring only about 10% of the time. Less than 8% of analyzed chatbots show large language model-level conversational capabilities, while more than 92% still rely mainly on rule-based systems, and roughly four in five bots do not clearly disclose their AI nature, creating both brand and looming regulatory risk.

In voice, Parloa’s own production AI voice agents were frequently blocked by rigid IVR trees and authentication flows built only for humans, forcing navigation through three to four menu layers, hold times exceeding 90 minutes, and offering little transparency or callback functionality. From Parloa’s perspective, this illustrates the incompatibility between legacy telephony infrastructure and emerging AI-driven service models, and underscores the commercial opportunity for enterprises that modernize around agentic, AI-ready architectures.

The most forward-looking concern flagged by Parloa is that only about 1% of tested enterprises are ready for agent-to-agent interactions in which a consumer’s personal AI agent talks directly to an enterprise’s service agent. Parloa argues that many brands are incorrectly treating AI callers as security threats because their systems cannot accept natural-language requests or authenticate software-based callers, a constraint the company frames as architectural rather than strategic and an impediment to future customer loyalty and revenue gains.

For Parloa, the study serves a dual role as a market diagnostic and a strategic sales asset, reinforcing its positioning as an agentic CX platform that can help enterprises rebuild infrastructure to support AI-driven, resolution-focused service at scale. The company plans to repeat the benchmarking annually and segment results by industry, a move that could deepen its data moat, sharpen its go-to-market messaging, and give executives at target clients a quantified view of their CX gaps relative to peers.

Founded by Malte Kosub and Stefan Ostwald, Parloa now employs about 430 people across New York, Berlin, London, and Munich and focuses on enabling global enterprises to design, train, and manage AI agents across voice and chat channels. The implications for stakeholders include near-term pressure on enterprises to upgrade outdated CX stacks, potential regulatory scrutiny of opaque chatbot disclosures, and expanded runway for Parloa to capture demand from brands seeking to improve service quality, reduce friction, and unlock higher customer lifetime value as AI agents become mainstream.