SKILLS AND TRAINING NEEDED
The migration to modern, AI-ready data platforms isn’t necessarily a brand-new phenomenon.
“The good news is that AI actually rewards the fundamentals,” said Arora. “For data professionals, this is a moment to double down on foundational hygiene and be brilliant at the basics. How clean your data is, how well your processes are documented, how consistently business rules are applied—all of it directly determines how useful your AI systems will be.”
Tried-and-true skills such as governance and cost management also are well adapted for the modern AI-ready platform, said Arora. “Understanding how AI consumes data, what it costs, and how to enforce responsible usage puts data professionals at the center of enterprise AI strategy, not on the sidelines of it.”
The fundamentals need a modern upgrade in certain areas as well. Today’s data managers need to “operate across infrastructure, analytics, automation, and business strategy,” said Ward. “Foundational database and integration skills remain important, but today’s leaders must also understand AI-enabled observability, predictive analytics, and workflow automation frameworks.”
This also requires an accurate understanding “of what happens when AI executes rather than recommends,” said Limburn. “Many data professionals still think about quality in terms of dashboard accuracy or model performance. Those remain relevant, but they underweight the consequences that come with agentic execution.”
For example, “When an agent triggers a financial workflow, initiates a compliance escalation, or routes a service decision without human review at each step, the organization is accountable for that action,” Limburn continued. “The data team that owned the underlying record is part of that accountability chain, whether or not they recognize it.”
This also throws a new light on data stewardship. “It shifts from a reactive practice of fixing problems when someone complains to an active practice of certifying data before any automated system is permitted to act on it,” Limburn said. This encompasses areas such as “quality scoring, identity resolution, business definition enforcement, and lineage completeness. Importantly as well, accountability for AI-driven decisions stays with the organization, not the technology vendor, which means the audit questions flow back to the data team eventually.”
Limburn also posed additional questions that need to be addressed in an AI-ready data platform environment. “If a dataset fails a quality threshold, what happens downstream? Which workflows are affected? What is the escalation path? That requires familiarity with both the data pipeline and the AI orchestration layer, a combination that most data roles have not historically needed. Building that cross-layer literacy is probably the most practically valuable investment a data professional can make right now.”
Those charged with deploying and maintaining AI-ready platforms “need stronger, programmatic engineering skills [along with] a strong understanding of Kubernetes-based environments, streaming data ingestion, and the ability to operationalize open table formats,” said Beauvais. “Skills in ML [machine learning] pipeline automation and agentic-AI orchestration are increasingly necessary as platforms host intelligent services that interact programmatically.”
Database DevOps also comes into play, especially with the growing complexity of today’s data platforms. “Professionals need to be skilled in version control, continuous integration, and automated releases,” said McMillan. “When it comes to AI and security, the key is learning how to safely adopt AI, especially as the technology brings forth a new variety of risks and challenges alongside the opportunities.”
THE ROAD AHEAD
Looking ahead within the next 3–4 years, “Modern data platforms will increasingly be judged by how well they continuously optimize the data estate, not just enable access to it,” Hiew said. “Expect more policy-driven automation, AI-assisted management, and platforms that treat modernization as an ongoing operational discipline, balancing agility, cost, and trust at scale.”
As modern, AI-ready data platforms evolve, expect to see an ongoing move “toward outcome-centric, intent-based platforms,” said Arora. “You’ll tell the platform what you want, and it will figure out how to deliver it.” There will be less emphasis on the “how—how data is moved, modeled, and prepared. Over the next 3–4 years, that will be abstracted away.”
Ultimately, the data platform of 2030, Arora predicted, “won’t be judged by how much data it manages. It will be judged by the trusted outcomes it drives. These systems will be purpose-built to produce the right information, in the right form, for the right agent or decision at the right moment.”
With the rise of AI as the most vital enterprise resource, “Data stacks are becoming the essential backbone for all enterprise AI initiatives,” said McMillan. “Architecture will be expected to support AI first and foremost. Organizations that spend the time to prepare their data for AI will see compounded benefits in the future.”