AI agents in enterprises: why PARLOA’s €310 million funding round marks a shift in the landscape

After two years of rapid expansion of artificial intelligence agents within enterprises, what was still perceived as an experimental field is gradually becoming a fully operational domain. Agents were initially deployed through more or less convincing proof-of-concepts. Capable of sustaining fluid conversations, processing simple requests, automating parts of customer support or internal assistance, they found their place in relatively controlled environments. While this phase allowed organizations to measure their potential, it also exposed their limitations.

An agent that works in a restricted framework does not necessarily guarantee robustness once integrated into the core of operations. When deployed along a customer journey, the agent no longer simply answers questions; it qualifies requests, redirects users, proposes actions, and sometimes executes decisions with significant consequences. Even when only partially autonomous, it becomes a sensitive component of the information system and, by extension, of the relationship between the company and its customers. At that level, errors are hardly acceptable.

From a promising tool to an operational asset

In many large organizations, discussions around AI agents have therefore gradually moved from the realm of innovation to that of governance.

During the initial phase, attention focused on the agents’ ability to understand intent, generate fluid responses, and reduce ticket volumes. While these indicators remain relevant, they become insufficient once the agent is expected to operate at scale.

Customer experience, IT, legal, security and risk departments are now questioning how to supervise a system that continuously learns and interacts with customers, sometimes in critical operational contexts.

Attention is concentrating on three essential dimensions. First, reproducibility, the ability to maintain consistent behavior across varied situations. Second, observability, which makes it possible to understand what the agent is doing, on what basis, and with what outcomes. Finally, correctability, which is essential to detect drift, adjust rules, test modifications, and redeploy them safely.

The agent therefore evolves from a simple tool into an operational asset and becomes subject to the same requirements as other components of the information system.

Customer experience as a laboratory for industrialization

If this maturation is particularly visible in customer service departments today, it is no coincidence. The high volume of interactions and the repetitive nature of many scenarios make CX a privileged observation ground. It also reveals the less visible costs of agentic AI. Automation does not eliminate complexity. It requires continuous human supervision, regular training cycles, exception management, integration with existing tools, and constant alignment with commercial policies and regulatory constraints.

Competition shifting toward the management layer

This issue has not escaped the attention of investors, who are increasingly interested in startups offering a control layer capable of turning agents into administrable systems rather than those promising the most fluid conversations.

Parloa illustrates this dynamic particularly well. The German company does not only emphasize the performance of its agents, but above all the capacity of its platform to design, simulate, supervise and evolve them within a controlled framework. This approach resonates with the expectations of large organizations confronted with reliability and accountability challenges.

2026: a year of normalization rather than disruption

Speaking of the end of “experimental” agents in 2026 does not mean the end of testing or innovation, but rather a shift in framework. Experimentation will increasingly take place within more formal structures, with validation criteria and supervision tools.

Companies will favor platforms capable of orchestrating several specialized agents, with the objective of evolving them coherently in line with operational constraints and strategic goals. In this perspective, the management layer becomes central.

Specialized players and emerging platforms

The competitive landscape of enterprise AI agents is structuring rapidly, though it has not yet stabilized. In Europe, the market remains fragmented between players specialized by verticals or use cases. In Germany, Parloa and Cognigy (acquired at the end of the year by NICE) hold positions in complex contact-center environments. In the United Kingdom, PolyAI has established itself in large-scale voice and conversational agents.

In France, the ecosystem remains more dispersed, with companies positioned on hybrid approaches, specific technological building blocks or sector-specific use cases, without a dominant platform standard yet emerging. In the United States, competition operates at a different scale, with startups such as Sierra alongside initiatives from major software and cloud vendors including Salesforce, Microsoft and Google, all progressively integrating agentic capabilities into their existing platforms.

A Series D round that changes Parloa’s scale

Founded in 2018 in Berlin by Malte Kosub and Stefan Ostwald, Parloa develops an AI agent management platform designed for large enterprises. The company employs around 380 people and operates offices in Berlin, Munich and New York. It reports working with several major international clients, including Allianz, Booking.com, SAP and Swiss Life.

Parloa recently announced a €310 million Series D funding round, led by General Catalyst with participation from EQT Ventures, Altimeter Capital, Durable Capital Partners and Mosaic Ventures. The transaction values the company at around $3 billion and brings its total funding to more than €480 million in less than four years.