Insights from Cisco’s Simon Miceli on overcoming complexity, building trust, and preparing for AI-driven business transformation
Frost & Sullivan Thought Leadership Series: AI Transformation
Artificial Intelligence (AI) is entering its most consequential transition—moving from experimentation to industrialization. As organizations accelerate their adoption of generative AI, the next major inflection point is emerging: agentic AI, where intelligence becomes autonomous, embedded, and deeply integrated into business processes.
Cisco’s AI Readiness Index, a global study of more than 8,000 business and IT leaders, underscores this urgency. Nearly seven in ten companies (69%) rank AI as a top IT budget priority, and within the next three years, almost all (86%) organizations expect AI use cases to deliver noticeable productivity improvements for employees. 📎 Cisco AI Readiness Index
To understand how enterprises can overcome this readiness gap, Frost & Sullivan spoke with Simon Miceli, Managing Director, Cloud and AI Infrastructure, Asia Pacific, Japan and China (APJC), Cisco during Cisco Live Melbourne.
We extend our sincere thanks to Simon Miceli and Cisco for their invaluable contributions to this series.
1. Why AI Matters Now: From Technology Shift to Business Transformation
AI today resembles the early days of the internet—full of potential, still complex, and on the verge of mass industrialization. Yet, as Simon explains, this transition may prove even more transformative:
“There are only a few technology transitions in our lifetime that fundamentally change everything. The internet was one. AI is another—and it is even more consequential.”
Unlike earlier cycles, AI’s impact is inherently business-first, not consumer-driven. This shift is defined by:
- Business-led transformation
AI is being deployed to improve customer experience, accelerate innovation, enhance workforce productivity, and streamline operations.
- The beginning of AI industrialization
Organizations are moving beyond pilots into genuine scale, building AI factories, data pipelines, integrated architectures, and secure-by-design systems.
- Universal conviction
Organizations understand that failing to embrace AI risks creating their next “Kodak moment.”
- A need to reduce complexity
AI infrastructure remains technically demanding. The goal is to free organizations from complexity so they can focus on business value.
Cisco’s long-standing engineering strengths position the company at the heart of this transition. As Simon notes:
“If we can take the complexity out of the infrastructure, we can democratize access to AI. The more accessible it becomes, the more time organizations can spend creating value.”
2. What’s Holding Adoption Back: Complexity, Cost, Trust, Data, and Use Case Maturity
Despite rising enthusiasm, Frost & Sullivan analysis finds that most organizations remain in the earliest stages of AI maturity. Simon’s observations across APJC highlight five dominant barriers.
- Complexity of Technology
AI today requires sophisticated technology. It requires:
- selecting and refining models
- orchestrating hybrid and multi-cloud environments
- securing data, pipelines, and endpoints
- integrating systems in production
Even advanced enterprises recognize that AI remains technically demanding.
- High Cost: Compute, Power, and Infrastructure Constraints
AI infrastructure is capital-intensive:
- exponential increases in power and cooling requirements
- GPU scarcity and long procurement lead times
- data center capacity limitations
- energy supply challenges across markets
Next-generation AI hardware will increase these demands further.
- Use Case Ideation and ROI
Organizations often know they need AI—but cannot articulate a clear business problem.
Simon notes:
“Many organizations think using copilots means they’re doing AI. But copilots are not business transformation.”
This leads to:
- unclear value propositions
- stalled projects
- weak ROI justification
- limited progression beyond pilots
The enterprises progressing fastest are those investing in structured use case ideation, commercial analysis, and well-defined proofs of concept.
- The Trust Gap
Trust determines AI adoption.
Organizations will not deploy AI systems they do not trust—from data privacy to model outputs. Concerns include:
- hallucinations
- lack of explainability
- data leakage
- safety and security gaps
- compliance risks
- insufficient governance
Simon emphasizes:
“If I can’t trust the intelligence I’m building, I cannot use it.”
This makes security-by-design essential.
- The Data Gap: The Most Significant Constraint
Data readiness remains the biggest barrier.
Most organizations struggle with:
- fragmented data ecosystems
- low-quality or unstructured data
- inconsistent governance
- limited data pipelines
As Simon notes:
“The model is only as good as the data made available to it. For many organizations, organizing their data is the biggest challenge.”
3. The Shift Toward Agentic AI: The Next Evolution in Enterprise Intelligence
While generative AI captures headlines, the industry is rapidly progressing toward agentic AI—where systems execute tasks autonomously and reason across complex workflows.
From simple one-to-one interactions…
(chatbots, copilots, prompt-based tasks)
…to autonomous, multi-step task execution and full business process transformation.
Agentic AI can unlock:
- autonomous decision-making
- workflow orchestration
- reasoning capabilities
- cross-system task execution
- dynamic adaptation to changing business conditions
As Simon explains:
“Agentic AI is what gives AI true business application. It’s what takes us from writing emails to fundamentally transforming business processes and customer interactions.”
Cisco’s strategy reflects this transition, with heavy investments in:
- Secure AI Factories, built from modular AI Pods
- Unified Edge Platforms for distributed inference
- deeply integrated networking and security
- teams of AI practitioners supporting customer ideation and ROI modeling
This marks the beginning of ubiquitous intelligence, where AI becomes pervasive across data centers, networks, and the edge.
4. Best Practices: What Organizations Should Do Next
Simon offers three clear, actionable recommendations aligned with Frost & Sullivan’s AI Transformation Framework.
- Remove Infrastructure Constraints
Organizations should:
- adopt reference architectures
- standardize deployment and operations
- ensure predictable access to high-density compute
- leverage modular AI infrastructures like AI Pods and AI Factories
- partner with providers that address power, cooling, and GPU availability
This reduces cost, complexity, and time-to-value.
- Don’t Bolt on Security—Build AI Secure-by-Design
Security must be embedded throughout:
- model training
- data pipelines
- inference and edge workloads
- identity and access controls
- monitoring, governance, and compliance
“If you cannot trust the results of the AI system, you cannot use it.”
Security is foundational—not optional.
- Prioritize Data Organization
Data readiness is essential for agentic AI.
Organizations must:
- establish strong data governance
- elevate data quality
- unify structured and unstructured data
- build consistent pipelines and metadata
- ensure data is usable and accessible for AI systems
AI maturity is impossible without data maturity.
Conclusion: Preparing for the Age of Ubiquitous Intelligence
AI is no longer a technology initiative—it is a business transformation mandate. The industry is shifting from productivity tools toward autonomous, agentic systems that will redefine how work gets done across every industry.
To prepare for this future, organizations must:
- Modernize infrastructure to unlock scalable, high-density compute.
- Embed secure-by-design principles to build trust and resilience.
- Prioritize data readiness as the foundation for intelligent automation.
The next era—the age of ubiquitous intelligence—has already begun. Organizations that build the foundation today will define the competitive landscape of tomorrow.
