Here’s a question I think worth pondering for customer experience (CX) leaders developing their artificial intelligence (AI) strategies and technology roadmaps: Is agent assist a permanent fixture of the modern contact center desktop, or simply a stopgap on the way to full automation?
It’s a question that’s been in my own thoughts over the last few weeks, initially landed there during a customer presentation at the NiCE Cognigy Nexus 2026 conference that took place March 11-12 in Munich. For global financial services firm Allianz Group, agent assist is a bridge technology that, while serving a useful immediate purpose, is perhaps even more valuable for its role in maturing the AI ecosystem. Such is the view relayed by Benno Schindler, conversational AI tribe lead at the company’s IT and digital arm, Allianz Technology, during his presentation, Allianz: Mastering Agentic AI and Building High-Impact Agents.
For Allianz, agent assist is the training ground for more advanced agentic AI — autonomous systems capable of making decisions and taking actions with little to no human involvement. Agent assist is necessary for refining agentic bots, Schindler said. By trying agentic bots in an assistive context first, the team gains the necessary experience and understanding to then eventually move them directly onto production lines.
Schindler outlines a clear migration path for agent assist technology at Allianz:
2025: Launch of transcription and summarization bots.
2026: Implementation of interventional agent assist.
2027: Transition to “background” assistance where the human guides the conversation on complex topics without being directly exposed to the interaction.
2028: Movement toward fully integrated systems where humans and bots are operated in agentic fashion.
While the long-term vision may be autonomous, the current benefits of agent assist are a challenge to quantify, in terms of dollars and centers, for Allianz, Schindler said. And soft metrics, like reduced average handle time or after-call work, aren’t going to help sway the budget holders, he added.
“A super good agent assist bot … does not decrease talk time a lot, maybe 10%. This will not pay for the new technology,” Schindler said. “After-call work production? That could pay, with summaries and ratings. Post call churn? That could pay. … But all of these are funny money. If you do a business case with funny money, your CFO will send you home until he sees real savings. It’s all great to argue, but it’s not going to win the case.”
Certainly, this is not to say that agent assist technology isn’t enabling companies to realize true benefits, as Metrigy knows from its CX research.
In fact, agent assist (sometimes called agent copilot) is the second-most-used AI application for CX, according to Metrigy’s most recent CX research, Customer Experience Optimization: 2025:26, a global study of 656 companies conducted in August-September 2025. More than 55% of companies already use agent assist to support customer interactions, and 39% are actively planning for implementation or evaluating use cases. What’s more, two-thirds of companies using agent assist reported improvements to agent quality, while 59% noted increased sales as a result.
Regarding increased sales, this is a trend that Metrigy has tracked for the last two years. As companies increase their use of agent assist, they see value for upsell and cross-sell. Agent assist doesn’t just help agents resolve customer issues; it gives them a way to bring revenue into the company.
It would seem that for companies like Allianz, finding the dotted line between agent assist and revenue increase will be imperative, especially when looking for budget approval.
The reality for Allianz is that agent assist may be a bridge in its current form, but its purpose — the tight integration of human intelligence and machine efficiency — is essential for the long haul. Schindler emphasized that the goal is for agent assist to be a tool for progress rather than a “self-fulfilling prophecy” that delays full automation.
The bottom line: Whether viewed as a lasting copilot or a bridge, agent assist technology can provide the context, process guidance, and — yes — return on investment necessary to survive the transition to an AI-first customer experience. It builds the infrastructure required for the autonomous world of 2028 while delivering measurable success today.