We can all agree that the AI revolution is here. But there is an emerging disconnect between ambition and readiness that is costing businesses millions in failed AI initiatives. In a recent webinar, we explored this challenge with industry experts Anthony Levesanos from Accenture Song and Phil Lockhart from Credera, uncovering the real requirements for building AI-ready data foundations that drive measurable business outcomes.
The Reality Check: AI Ambition vs. Data Readiness
“The bigger blocker right now isn’t the AI tech, it’s that most brands don’t have clean, connected, consented data,” explains Anthony Levesanos, Managing Director and Head of Data & AI at Accenture Song. “The technology is moving faster than most companies’ data strategies can keep up with.”
This reality is playing out across industries. Consider Klarna’s highly publicized AI customer service initiative that replaced 700 employees with AI tools, only to rehire them when the AI solution failed to meet quality expectations. The issue wasn’t the AI capability—it was the foundation supporting it.
Contrast this with success stories like EasyJet, which uses AI to create highly personalized travel itineraries that help customers save time and discover trips aligned with their preferences. The difference? A solid data foundation that enables AI to work with context, not just algorithms.
Rethinking First-Party Data: From Collection to Connection
“It’s not what you collect—it’s what you connect,” says Phil Lockhart, Chief Digital Officer at Credera. This shift in thinking represents a fundamental change in how organizations approach their data strategy.
Historically, companies treated first-party data like oil—a raw asset to collect and store. But in an AI-powered world, data functions more like embedded GPS. When your data foundation is missing or disconnected, AI doesn’t stop—it keeps going, just not where you want it to go.
The Three Pillars of AI-Ready Data
According to our expert panel, successful AI implementations require three foundational elements:
- Strong Data Foundation
- Unified data platforms that consolidate sources
- Advanced processing tools for large-scale cleansing and transformation
- Robust governance embedded in system architecture
- Flexibility
- Adaptable data layers that evolve with changing business logic
- Model-agnostic architectures that can incorporate new AI capabilities
- Real-time responsiveness to changing consent laws and customer behaviors
- Embedded AI Functionality
- Decision-making capabilities built into workflows, not bolted on afterward
- Real-time processing that acts on current behavior, not last week’s dashboard
- Automated guardrails that ensure compliance while enabling innovation
The Privacy-Innovation Balance: Building Trust Through Design
“Just because you’re collecting the data does not mean that you can use the data,” began Lockhart. “AI agents must make contextual, real-time decisions rather than relying on generic or outdated responses, since they’re operating continuously. Tealium demonstrates this perfectly. Their platform streams event-level data with built-in consent controls, enabling agents to make decisions that are both immediate and compliant with regulations.”
A common misconception is that privacy and compliance requirements slow AI innovation. Our experts argue the opposite: privacy concerns only become innovation blockers when treated as afterthoughts.
“The smartest companies build privacy into their processes and systems from the start,” noted Lockhart. “When privacy is wired into the architecture and not slapped on later, it accelerates innovation by reducing risk, speeding approvals, and minimizing rework.”
Levesanos added, “Governance is non-negotiable. It’s foundational for AI viability across every industry. With clear governance you can scale AI in a responsible manner.”
This approach requires three key components:
- Consent-aware data flows that power compliant personalization
- Audit trails built into AI decision-making from the ground up
- Guardrails that enable experimentation without crossing legal boundaries
The Rise of Agentic AI
The next frontier is agentic AI, which represents systems that can reason, decide, and act with minimal human supervision. These aren’t just enhanced workflows; they’re autonomous agents that own entire processes.
“Agentic AI raises the bar for readiness in a big way,” explained Levesanos. “It’s not just about clean data anymore—it’s about connected systems, real-time orchestration, embedded governance, and human oversight working together.”
Lockhart also noted, “The problem is that oftentimes the guardrails are very wide on agentic solutions, and it gets off track. And, when it gets off track, guess what? It’s off track 24/7 and doing things you don’t intend it to do. But when you can identify a very clear lane and focus, you can build multiple agents that work together to complete different tasks, and then have another agent that directs and compounds all these things together.”
This evolution requires three major architectural shifts:
- Modular, Event-Driven Design: Systems that respond dynamically to real-time events and context
- Embedded Governance: Real-time policy checks, guardrails, and explainability built into the foundation
- Human-AI Collaboration: Rethinking roles and trust between humans and machines
Getting Started to Future-Proof Your Business
For organizations beginning their AI journey, success comes from balancing quick wins with foundational investments.
“As agentic AI and any advanced tech becomes ubiquitous, it’s going to become a commodity. Creativity: how we solve problems, design, communicate, and connect will become the true differentiator,” said Levesanos. “The technology might be the same across the board, but what you do with it, that’s where future proofing really happens.”
Here’s what you can do to get started:
Prove Value Quickly: Start with high-impact use cases that solve costly, inefficient, or customer-critical problems. Work backward from the business problem to identify exactly what data you need.
Build for Scale: Simultaneously invest in data governance platforms, talent development, and modular architecture that enables rapid scaling when initiatives prove successful.
The key is focusing on specific, relevant datasets that are “good enough” for targeted use cases rather than pursuing comprehensive data perfection across everything.
Success in this new era demands the cleanest, most connected, most compliant data flowing through systems at every moment that matters. The future isn’t evenly distributed yet, but it’s arriving faster than most organizations are prepared for.
For more insights, check out the webinar, “Building a Scalable AI-Ready Data Strategy with Accenture, Credera, and Tealium.”
