The rise of autonomous artificial intelligence (AI) — systems that can make decisions and act autonomously — represents a change in how organizations shift to a new paradigm for business operations.
Unlike traditional AI tools that respond to prompts, AI agents can reason, plan, and execute tasks independently, which makes them a powerful force for transformation. But IT leaders must navigate their agentic AI journey carefully to ensure successful adoption.
That’s why Foundry reached out to the CIO Experts Network — a community of IT professionals and technology industry influencers — and asked them how IT leaders can set up their organizations for success.
“Executives often misunderstand agentic AI,” says Peter Nichol, data and analytics leader for North America at Nestlé Health Science. “It’s not an intern who never sleeps, and it’s not about making random decisions. Agentic AI is a focused, reliable teammate.”
That said, it’s crucial to have visibility into and control of agents because they enable organizations to innovate and experiment with confidence. Scott Schober, president and CEO at Berkeley Varitronics Systems recommends thinking about agentic AI “as a new employee or new team member. Proceed with caution, set tight boundaries, and observe everything it does very closely.”
William Benjamin, principal generative AI and machine learning expert at Allstate, emphasizes agents’ ability to solve complex problems on their own. “Agentic AI isn’t just smarter chatbots. It’s systems that think and act independently, making decisions in real time. This autonomy excels in fast-paced scenarios like troubleshooting outages or optimizing supply chains, where delays can be costly.”
Build the foundation: Data and infrastructure excellence
Before deploying agentic AI, organizations must establish a robust foundation. Gene de Libero, principal at marketing technology consultancy Digital Mindshare, cuts straight to the essentials: “Get your tech basics right first: clean APIs and good data.”
Enterprises should also format data so that agentic AI can fully understand it contextually. Michael Bertha, partner at Metis Strategy, points to a more nuanced requirement: “Real value comes when [agents] understand what the data means, semantically and operationally, within the business process.”
“Organizations are centralizing knowledge repositories and cleansing data stores that can be used in private language models,” adds Isaac Sacolick, president of StarCIO and bestselling author. “In parallel, IT leaders are updating business workflows with reliable integrations and robust exception handling.”
In addition to a strong data foundation, ensure you’re on solid ground regarding security and governed access because agentic AI will act independently. “The only way true success can be achieved is through well-defined use cases, restricted access to sensitive systems, and robust human supervision from the outset,” Schober says. “Build security and transparency into your systems from the beginning.”
Start small and expand
Agentic AI has enormous potential to transform key operational functions and increase efficiency, but deployment doesn’t require an immediate organizational overhaul. Instead, experts recommend a measured, iterative approach. Start small, establish a beachhead that demonstrates clear value, and then expand. Over time, enterprises will need to reimagine their business processes, but it will happen gradually.
De Libero suggests having a tight focus: “Pick two to three real business problems, not random pilots everywhere,” he says. “Build the foundation, prepare your team, and focus on what actually makes money. It’s about workplace transformation and changing how you work, not just buying AI tools.”
Of course, choosing the initial project is no simple matter, given the vast array of potential use cases. Nichol advocates for quick wins that build early confidence. “A self-healing dashboard might auto-fix broken data sources or refresh jobs before users file tickets,” he says. “A code archaeologist could crawl GitHub to flag dead code and orphaned pipelines.”
Once the use case is identified, provide IT a safe sandbox for experimentation. “Pilot agentic AI in controlled environments using containerized architectures, allowing your teams to safely explore autonomy without overhauling existing infrastructure,” says Will Kelly, a writer who focuses on AI and the cloud. “Lean into tools like custom GPTs or internal AI playbooks to speed up experimentation and ensure consistency in how AI-generated outputs are validated.”
Kelly also emphasizes the importance of working collaboratively across different functions. “Adopting agentic AI requires IT leaders to blend innovation with solid operational grounding,” he says. “Start by establishing a cross-functional foundation that includes your devops and cloud teams since their familiarity with containerization and orchestration will ease integration with AI workloads.”
Get people onboard
Organizations have placed a high priority on implementing AI, but that doesn’t mean there’s always enthusiastic support for an agentic AI endeavor. Many C-suite executives will be favorably inclined towards agentic AI, but must still be convinced of the business benefits. Obtaining their signoff, as well as their participation, will be critical for long-term success and the ability to scale.
“Secure C-suite sponsorship to align AI deployment with core business goals, delivering outcomes like faster operations and a competitive advantage,” Benjamin says.
In other words, tie AI initiatives to strategic business goals and KPIs. For example, Nichol suggests combining “hard ROI — cost savings, revenue lift — with soft ROI — speed, satisfaction, compliance — to build a full-spectrum business case.”
Also, “monitor indicators like time reductions, productivity increases, and shifts in strategy, leveraging prototypes for rapid feedback,” Benjamin says.
In addition, enterprises should prioritize engagement and adoption among the employees who will work with agentic AI on a day-to-day basis. Training is important, as well as user-friendly tools, such as low code or generative AI capabilities that are built into products. Doing so democratizes and accelerates innovation.
Agentic AI represents a significant opportunity if approached strategically. Success requires more than implementing new technology. It demands a fundamental shift in how organizations think about existing enterprise processes, decision-making, and human-AI collaboration.
Start small with a solid foundation, maintain trust and governance, and focus on real business value. By following these principles and learning from implementations, IT leaders can harness the full potential of agentic AI while managing its inherent risks and complexities.
Start building your agent-ready foundation with the expert guidance of MuleSoft, which has extensive experience helping enterprises identify, deploy, and adopt agentic AI use cases.