AI agents move from chatbot toy to enterprise co‑worker
Enterprise AI agents are software systems that combine large language models, workflow logic and integrations to autonomously handle complex business tasks such as customer conversations, supply chain decisions and marketing execution, going far beyond simple chatbots by taking actions, following policies and continuously improving from feedback. The surge of AI agent funding is not an experiment anymore; it is a clear bet that the next generation of enterprise automation will be agentic. Legaltech startup JUPUS raised €13 million in a Series A to expand what it calls an AI secretarial service for law firms, automating client calls, case preparation and document drafting. Berlin-based industrial AI startup Kyrok secured €3.1 million in pre-seed funding to build an AI operating system for pharmaceutical and chemical supply chains powered by industry-specific AI agents. These are not side projects. They are early signs that investors see agents as the new systems of record for work.
Vertical AI solutions beat generic platforms to budgets
The most telling pattern in recent AI agent funding is how vertical AI solutions have elbowed past broad, horizontal tools. JUPUS has built what it describes as Europe’s first AI secretarial service purpose-built for law firms, targeting a 70 percent decline in newly trained legal assistants over three decades even as the number of practising lawyers has tripled and administrative workloads have grown. By automating client calls, inquiry triage, case structuring and document drafting, JUPUS claims to save firms more than 70 hours per month and already processes over 2,000 new legal cases per day for more than 2,000 lawyers. Kyrok, meanwhile, sits directly on top of ERP systems to codify pharmaceutical and chemical supply chain know-how into agents that guide order intake today and aim to expand into production planning, material planning and procurement tomorrow. These agentic workflow platforms win because they understand the messy, regulated workflows of their niche better than any generic assistant ever could.
Marketing and customer engagement agents show the loudest signal
If you want to see where AI agents are already compounding revenue, look at marketing and customer engagement. Gradial raised USD 65 million (approx. RM299.5 million) in a Series C to build a “system of work” for enterprise marketing, pairing AI workflow agents with content infrastructure that stores brand, content and process context so teams can cut campaign execution times by more than 80 percent while keeping brand and accessibility compliance intact. Respond.io closed a USD 62.5 million (approx. RM287.5 million) Series B as its AI-powered messaging platform hit USD 35 million (approx. RM161 million) in annual recurring revenue, growing 169 percent year-over-year at a 30 percent profit margin while handling two billion messages per quarter. Its AI agents qualify leads, answer customer questions and close sales autonomously, charging by conversation volume rather than seats so automation increases efficiency without shrinking revenue. JustAI’s USD 17 million (approx. RM78.2 million) Series A and 5x ARR growth round out the picture: agentic marketing platforms that coordinate strategy, creative, decisioning and data into one continuous decisioning layer are becoming the default ambition for modern marketing orgs.
High-touch workflows are where enterprise automation startups win
The unifying thread across these enterprise automation startups is their focus on high-touch workflows that once depended on scarce humans. In law, JUPUS steps in where legal assistants are missing, handling client calls, structuring inquiries, preparing cases and drafting documents so lawyers can focus on legal strategy rather than inbox triage. In pharma and chemicals, Kyrok’s agents wrap around existing ERP systems, capturing tacit knowledge from aging supply chain experts and turning it into guided workflows for order intake today and, soon, production, material planning and procurement. On the customer front, Respond.io’s agents orchestrate millions of conversations across WhatsApp, Instagram, TikTok, Telegram and WeChat, resolving support requests and closing sales autonomously. Voice is becoming another high-touch frontier: enterprises are adopting voice AI for customer service, sales, financial services and healthcare, and Coval’s testing and monitoring infrastructure is what lets them deploy autonomous voice agents with confidence at scale. Enterprise buyers are voting with their budgets for agents that remove the bottlenecks in scheduling, procurement, voice interactions and customer conversations.
Agentic workflow platforms signal the next wave of enterprise software
The funding wave is not about chat widgets; it is about agentic workflow platforms that make autonomous decisions inside enterprise constraints. Gradial frames its product as AI agents executing workflows on top of a cloud content infrastructure that stores brand and process context, a move from rules-based automation to goal-oriented systems that reason, act and learn. JustAI explicitly presents its four-agent model—strategy, creative, decisioning and data—as infrastructure for autonomous decisioning, designed to consolidate audience analysis, creative production, decisioning and measurement that used to be scattered across tools. Coval, meanwhile, is building the evaluation and observability layer for voice and chat agents, running tens of millions of simulations so enterprises can move from pilots to reliable production, and it is doing so in a voice AI market that saw more than USD 7 billion (approx. RM32.2 billion) invested in the first quarter alone and is expected to reach USD 20 billion (approx. RM92 billion) by 2031. In other words, investors are treating agentic automation as the next operating stack for work.
The takeaway for buyers and builders is blunt: the window for “AI-powered” labels without agentic depth is closing. Workflows that involve repetitive, high-volume decisions with clear constraints—like marketing execution, customer messaging, legal intake and supply chain coordination—are being redefined as continuous, autonomous systems. The winners will be those who pick or build vertical AI solutions that own the workflow, not those who add yet another chatbot to an already crowded tech stack.
