Scaling agentic AI pilots across the enterprise
As agentic AI moves from experimentation toward enterprise deployment, the challenge is figuring out how agents can work together, connect to the systems and data they need, and operate safely across the workflows that run a business. Although agentic AI has been adopted by some 80% of Fortune 500 companies, progress toward meaningful scale remains….
Agentic AI refers to artificial intelligence systems designed to act autonomously, make decisions, and perform tasks without constant human oversight. Unlike traditional AI, which follows predefined rules, agentic AI can adapt, learn, and interact with other systems or agents. Currently, around 80% of Fortune 500 companies have adopted agentic AI, but many are still in the early stages, running small-scale experiments rather than deploying the technology widely. The challenge lies in moving beyond isolated pilots to integrate agentic AI into core business operations. Organizations must first define clear business goals, such as increasing revenue or reducing costs, before scaling the technology. Simply adding AI to existing workflows without redesigning those workflows risks inefficiency, as outdated processes may not benefit from AI advancements. Arun Chandra, Chief Operating Officer at NiCE, emphasizes that agentic AI should be treated as a cohesive system, not a patchwork of individual tools.
To scale agentic AI effectively, organizations must ensure that AI agents have access to the right data, knowledge, and context to perform their tasks. Without high-quality, relevant information, AI agents cannot make informed decisions or take meaningful actions. Fragmented or siloed data can severely limit their effectiveness, as Chandra notes: the efficacy of AI agents depends entirely on the quality of the context and data they can use. Additionally, agents must be connected to back-end systems, such as databases or enterprise software, to execute tasks like processing orders or updating records. Beyond technical requirements, scaling agentic AI also involves addressing organizational challenges. Teams must avoid creating isolated AI systems that don’t communicate with one another, as this can lead to inefficiencies. Strong governance, privacy protections, security measures, and change management strategies are essential to ensure AI agents operate safely and align with business objectives.
Arun Chandra argues that AI agents should be held to the same standards as human employees, meaning they should be reliable, accountable, and integrated into the broader workforce. This approach treats AI not as a separate tool but as a complementary part of the organization’s human capital. For example, AI agents could handle routine tasks like scheduling meetings or analyzing data, while humans focus on strategic decision-making or customer interactions. The goal is to create a hybrid workforce where AI and humans collaborate seamlessly. This perspective shifts the focus from experimenting with AI to systematically integrating it into workflows, ensuring it enhances productivity and efficiency rather than disrupting existing processes.
In the future, agentic AI systems could become more proactive and collaborative, working together to resolve complex tasks. For instance, an AI agent handling a customer complaint might automatically communicate with another agent responsible for inventory management to check product availability or with a logistics agent to arrange a return. This interconnected approach could streamline operations and improve customer experiences. However, organizations should avoid attempting to scale AI across all areas at once, a challenge Chandra describes as “boiling the ocean.” Instead, businesses should focus on high-value use cases with measurable outcomes, gradually building a connected AI strategy that aligns with their workflows, workforce changes, and strategic goals.
This article is a sponsored piece produced in partnership with NiCE, a company specializing in AI solutions. The content was created by Insights, MIT Technology Review’s custom content division, not by its editorial team. The research and writing were conducted by humans, with AI tools used only for production processes under human oversight. The article highlights the challenges and opportunities of scaling agentic AI in enterprises, emphasizing the need for strategic planning, data integration, and workforce alignment. It also includes references to unrelated topics, such as vulnerabilities in large language models (LLMs) and ethical concerns in AI hiring, to provide broader context on AI advancements and risks.
