AI Expo 2026 Day 1: Empowering the Agentic Enterprise Through Governance and Data Readiness
As we step into the future of technology, the integration of artificial intelligence into everyday business practices is becoming more pronounced, especially in spaces like the AI & Big Data Expo and the Intelligent Automation Conference. Day one of these events highlighted the transformation from traditional automation to sophisticated, autonomous systems that not only follow instructions but also make decisions. This evolution promises to redefine how businesses operate, improving efficiency and productivity in ways we never thought possible.
The Shift to Agentic Systems
A focal point of discussion at the conference was how we are transitioning from passive automation to what are known as “agentic systems.” These advanced tools don’t just execute pre-defined tasks; they also reason, plan, and adapt in real-time. Amal Makwana from Citi shed light on this paradigm shift, demonstrating how these systems enhance workflows across different enterprises, setting themselves apart from outdated Robotic Process Automation (RPA).
Scott Ivell and Ire Adewolu of DeepL elaborated on closing the so-called “automation gap.” They argue that agentic AI acts as a digital co-worker rather than a mere tool, unlocking true value by minimizing the gap between intention and execution. Brian Halpin from SS&C Blue Prism echoed this sentiment, emphasizing the need for organizations to master standard automation before diving into deeper implementations of agentic AI.
Establishing Governance Frameworks
This evolution does not come without its challenges. There is a pressing need for governance frameworks to manage the non-deterministic outcomes associated with these intelligent systems. Steve Holyer of Informatica, along with representatives from MuleSoft and Salesforce, stressed the importance of rigorous oversight in architecting these operations. A governance layer is essential to ensure that agents can access and efficiently utilize data, thereby preventing potential operational failures.
Data Quality: The Foundation of AI Success
Quality input is critical for autonomous systems, a point underscored by Andreas Krause from SAP. Without trusted, connected enterprise data, any AI initiative is likely to falter. For Generative AI to be effective in a business context, it must engage with accurate and relevant data.
Meni Meller from Gigaspaces tackled the technical hurdle of “hallucinations” in large language models (LLMs). He advocated for a strategy that includes eRAG (retrieval-augmented generation) along with semantic layers to address data access challenges. This innovative approach enables models to tap into factual enterprise data in real time.
Moreover, the significance of cloud-native, real-time analytics surfaced during a panel discussion featuring industry leaders from Equifax, British Gas, and Centrica. For these organizations, staying competitive hinges on their ability to execute scalable and immediate analytics strategies.
Safety and Observability in AI Deployment
The incorporation of AI extends into physical environments, raising unique safety concerns that differ from traditional software-related issues. A panel featuring Edith-Clare Hall from ARIA and Matthew Howard from IEEE RAS discussed the deployment of embodied AI in various settings, including factories and public spaces. Establishing safety protocols prior to human-robot interactions is an absolute necessity.
Perla Maiolino from the Oxford Robotics Institute introduced a technical perspective on this challenge. Her research focuses on Time-of-Flight (ToF) sensors and electronic skin technology designed to provide robots with self-awareness and awareness of their surroundings, a crucial aspect for industries such as manufacturing and logistics to prevent accidents.
In the realm of software development, the issue of observability is equally important. Yulia Samoylova from Datadog highlighted that as AI becomes more autonomous, the ability to monitor internal states and reasoning processes is vital for ensuring system reliability.
Overcoming Infrastructure and Cultural Barriers
Successful implementation of these systems requires not just advanced technology but also a receptive cultural landscape. Julian Skeels from Expereo pointed out that networks must be tailored specifically for AI workloads. This involves constructing secure, “always-on” network fabrics capable of high throughput.
However, human unpredictability often complicates the adoption process. Paul Fermor from IBM Automation warned about the "illusion of AI readiness," noting that traditional automation approaches may underestimate the complexities involved in adopting AI. Jena Miller reinforced this, stating that human-centered strategies are paramount for genuine adoption. If employees lack trust in these tools, the technology is unlikely to yield a positive return.
To facilitate success, Ravi Jay from Sanofi suggested that leaders engage with both operational and ethical considerations early in the process. Key decisions revolve around when to develop proprietary solutions versus opting for established platforms.
Conclusion: Building a Strong Future Together
As we navigate these transformative changes, it’s evident that moving toward autonomous agents requires a solid data foundation. CIOs should prioritize establishing robust data governance frameworks to support retrieval-augmented generation, ensuring that network infrastructures can accommodate the demands of agentic workloads. Concurrently, cultural adoption strategies must align alongside technical implementations to maximize effectivity.
Together, we can leap into this exciting new era of AI, where innovation and collaboration define our future. Embrace this journey, and let’s shape a smarter, more efficient world together!
