Enterprise AI

Chairpersons welcome and opening remarks

As organizations mature their data platforms, the focus is shifting from simply storing data to delivering trusted, reusable data products that drive real business outcomes. This panel explores the evolution beyond traditional data lakes, examining data mesh, product-centric ownership, governance at scale, and the technologies enabling faster, more reliable analytics and AI. Industry leaders will share practical lessons on what’s working today, common pitfalls to avoid, and how enterprises can design their next-generation data ecosystems for agility, value, and impact.

As Head of the Global AI Delivery Unit at NTT DATA, I will share what we have learned from delivering AI at scale across industries and how organizations can create value today while preparing for an AI-native future.

The biggest mistake organizations can make with AI is waiting for perfect data readiness. The second is deploying isolated use cases without the architecture and operating model needed to scale.

Successful AI transformation requires more than data and models. It demands a foundation that combines governance, business alignment, talent, and a scalable AI architecture. Through our AI Delivery Unit, we have learned that organizations accelerate AI adoption by standardizing how AI is delivered – not by waiting until every dataset is perfect.

Most organisations don’t struggle to start with AI. They struggle to make it stick. A central team runs pilots, the business watches, and little changes in how work gets done. Becoming AI-native means AI is built into the daily work of every function, not kept in a lab.

This session shares a practical hub-and-spoke model for getting there without a large AI team. A small AI and data hub sets the standards, platforms and guardrails. The spokes are embedded partnerships with operations, risk, compliance, finance and engineering, where the real problems and the real value sit.

Drawing on real implementation experience, the session covers:

• Using a maturity framework to baseline each team and set clear targets
• Running structured discovery to uncover pain points and map processes
• Turning discovery into a 12-month roadmap that balances quick wins with lasting change
• Taking AI from idea to production in areas like operations, compliance and software delivery
• Building governance and trust so AI adoption scales safely
• Attendees will leave with a repeatable playbook for moving their organisation from isolated pilots to AI-native ways of working.

A faulty application change often fails fast. A faulty data pipeline can succeed, write incorrect data, and quietly feed products, dashboards, or AI training until someone notices downstream.

At Datadog, batch data pipelines increasingly power analytical products and AI model training. Their authors are often product engineers, AI researchers, and analysts rather than data specialists. We are making AI coding agents the primary interface for data development. They help engineers move from intent through pipeline design, implementation, and testing, then diagnose failures when they occur.

Generating pipeline code is becoming easier. Trusting it is not. This forced us to rethink how agents test against production data without risking production, distinguish an intended change from a data quality issue, and investigate failures across the stack. This talk follows that journey through our Airflow and Spark platform.

We will share what works today, where autonomous workflows still require human review, and what we learned while opening data engineering to more engineers.

Your marketing agent sees stock in the warehouse and offers next-day delivery to every customer on your list. Your sales agent starts confirming the orders. But your supply chain agent has already reserved that same stock for your biggest account last week. Each agent did its job. By morning, thousands of customers hold a promise you can’t keep.

As AI moves from answering questions to taking action, the risk is no longer a wrong answer. Today, it’s your own agents working against each other in your company’s name, at machine speed. Soon, your customers will send their own agents to buy from yours, and nobody on either side will be there to catch the mistake. Whatever your agent promises, your company owes.

Snowflake’s AI control plane doesn’t replace the tools your teams use to build agents. It sits underneath them. Every agent works from the same governed data and business definitions, so your marketing agent sees that the stock is already promised, and your customer’s agent gets one answer from your company, not three. Each agent gets only the access its task needs, reaches models and tools through one governed route, and leaves a record of what it did and on whose behalf. Your teams keep the agents and models they prefer, and every new agent builds on the business knowledge you already trust, instead of starting over.

You’ll leave knowing how to give AI more responsibility, with confidence.

How do you turn master data from an IT concept into a true business enabler in a global organisation like Shell?

In this session, Nathalia Beekman shares her journey of implementing Master Data Management at scale, navigating complexity, breaking silos, and creating a single foundation of trusted data across the enterprise. We’ll unpack the real challenges large organisations face, why MDM is often seen as the “unicorn,” and how we made it tangible and valuable.

You’ll hear practical lessons learned, proven best practices, and how we drive adoption by aligning IT and business perspectives so master data doesn’t just exist, but actively powers better decisions, efficiency, and impact across Shell.

Handing your engineers an AI copilot is easy. Getting your whole organisation to build better software with AI is the hard part. Once everyone is fast, the bottleneck just moves: to your product managers, your reviewers, your release cycle. The gap between your most and least fluent people widens instead of closing, and the output quietly drifts towards generic slop. Speed on its own is not the win.

This talk is the practical, under-the-hood story of how In The Pocket built an agentic software factory: 400 production skills, chained into workflows that run across the entire software development lifecycle, from research and engineering to quality and release. This is the real engine behind how we build today, and how it turned quality from our biggest risk into our biggest advantage. No theory and no hype, just the walls we hit and the system that got us through them.

Expect to leave with things you can apply when you’re back at the office. We’ll get practical about how to roll this out across teams and not just individuals, how to govern hundreds of moving parts with clear ownership and automated quality checks, how to close the gap between your most and least fluent people, and how to keep humans accountable for everything that ships. You’ll walk away with a real playbook for scaling AI-driven development across your organisation, not just another demo.
Building the right data pipelines, monitoring systems, and collaboration frameworks for sustained business value.

Cultural and technical enablers for scaling data-driven decision-making globally.

AI is no longer future technologies, it is reshaping how enterprises operate, compete, and grow. Yet many organizations remain stuck between promising pilots and real business impact.

This session explores how enterprises can successfully transform by embedding AI into their core strategy, operations, and culture. Moving beyond hype, we will look at what truly differentiates organizations that scale AI from those that stall: leadership alignment, data maturity, cross-functional collaboration, and a relentless focus on value.

Through real-world examples and practical frameworks, this talk will help leaders understand how to turn AI into sustainable transformation engines driving efficiency, innovation, and long-term resilience.

What makes a location the right place to grow a technology business in Korea?

This presentation introduces Suseong Alpha City, a rapidly growing digital innovation hub in Daegu, home to more than 300 companies and a growing ecosystem of AI, ICT and digital technology businesses.

Rather than focusing solely on incentives, the session will explore what makes Suseong Alpha City a practical base for global technology companies — from its location and business infrastructure to its expanding digital ecosystem, talent pool, research and support institutions, and connections to the wider Daegu-Korea innovation network.

The presentation will also highlight the ongoing expansion of Suseong Alpha City and the support available to companies looking to enter, establish and grow their business in Korea.

Discover why Suseong Alpha City could be your next gateway to the Korean market.

This use case explores how IKEA is using GenAI-powered assistants and agents to enhance customer experience across digital and physical touchpoints while keeping human judgment at the center. From personalized product recommendations to post-purchase support, AI systems augment co-workers rather than replace them, helping customers navigate complex choices with confidence.

As conversational agents become more autonomous, anticipating needs, resolving issues, and optimizing interactions purposeful leadership ensures transparency, brand alignment, and trust. Human oversight remains critical in defining tone, escalation paths, and ethical boundaries.

Attendees will gain insight into how IKEA balances automation with AI into omnichannel journeys, and empowers co-workers to supervise, refine, and continuously improve AI-driven customer interactions.

How do organisations turn data and AI into measurable business outcomes?

This discussion brings together product, platform, and business leaders to explore operating models that align data foundations with AI goals.

Case studies will show how cross-functional teams bridge silos, balance agility with governance, and deliver ROI from AI initiatives.