AI Frontier Post
AI News

TELUS Digital partners with Parloa on AI agents for contact centers — fewer than 9% of chatbot chats ever get resolved

TELUS Digital announced a partnership with Parloa on Wednesday, taking on an implementation-partner role for the German company’s agentic AI platform for enterprise customer experience. TELUS Digital will introduce enterprise clients to Parloa’s platform and help move it into production at scale — while enterprises contract directly with Parloa for the platform itself and engage TELUS Digital for implementation, integration, change management, and managed services.

The deal

Parloa’s agents are built to run a customer interaction end to end — authentication, bookings, payments, order inquiries — and hand off to a human agent, with full context, when a conversation calls for judgment. The platform runs on Microsoft Azure and supports multiple AI models, including OpenAI’s GPT series, and Parloa says it is built to meet compliance standards including GDPR, ISO 27001 and SOC 2. It also recently earned SAP’s endorsed app premium certification, a category SAP reserves for solutions meeting added security, testing, and performance standards.

For TELUS Digital, the partnership extends an existing AI-for-CX stack: conversational AI combining voice, generative AI and agentic automation, plus agent-assist tooling — Fuel iX Agent Assist gives agents real-time knowledge and answers, and Fuel iX Agent Trainer uses AI-generated voice and chat simulations to rehearse new agents before they take live calls. The company backs it with a global delivery footprint of 83,000+ people across 35+ countries.

Customer service agent wearing a headset
Where the AI meets the human: Parloa’s agents resolve eligible interactions end to end and escalate with full context when judgment is needed. Image: TELUS Digital.

The 9% problem

The announcement leans on Parloa’s own numbers for why agentic CX is more than a chatbot relaunch. A study of 10,000 enterprise websites and 4,000 chat interactions found that fewer than 9% of chatbot conversations reached resolution — and only about 10% of escalation attempts actually reached a human agent. Separately, Parloa’s Consumer Patience Index, a survey of 1,001 U.S. consumers by Propeller Insights, found that nearly a third of consumers would switch brands entirely just to avoid being put on hold.

“The research confirms what we’ve all experienced ourselves — consumer expectations are high, and brand loyalty is fragile,” said Chris Silver, Parloa’s Chief Revenue Officer. The pitch is that completing an interaction, rather than stalling partway through it, turns automation into loyalty instead of frustration. “When contact centers get buried in high volume, repetitive interactions, we often see cost to serve climb without the customer experience improving,” said Will Mayo, Global Senior Vice President of Growth at TELUS Digital.

Compliance as the moat

Parloa, founded by Malte Kosub and Stefan Ostwald, employs 500+ people across eight offices in the U.S. and Europe, and says it has six years of deployments in banking and financial services, insurance, utilities, telecommunications, media, retail, logistics, travel, and healthcare and life sciences. The geographic pairing is deliberate: Parloa brings a European enterprise footprint and GDPR-seasoned deployments, while TELUS Digital brings delivery presence and regulated-industry experience across North America and Europe.

Agent with headset in a customer support center
Implementation is the product here: TELUS Digital will act as a one-stop shop from strategy to scaled delivery of Parloa’s platform. Image: TELUS Digital.

Why it matters

The move lands in a contact-center market where the limiting factor on AI automation has shifted from model quality to operational trust — security review, compliance sign-off, and someone to own the transformation when the rollout stalls. Partnerships like this are how agentic AI actually gets deployed at enterprise scale: one vendor owns the platform, another owns the services muscle to land it. The benchmark to watch is simple: how many interactions get resolved end to end, and whether the human agents left standing get the harder, higher-value conversations.