Strategic Priorities for Data and AI Leaders in 2025
Summary
As businesses across industries prepare for next year’s AI investments and strategies, we share where organizations are focused for 2025:
- Infrastructure will be the most urgent investment as companies build AI agents and systems
- Enterprises laser in on their data advantage to use AI for a competitive edge
- AI training will focus on workforce behavior changes to increase adoption and build human-in-the-loop processes
AI remains at the forefront of every business leader’s plans for 2025. Overall, 70% of businesses continue to believe AI is critical to their long-term success, according to a recent survey of 1,100 technologists and 28 CIOs from Economist Impact. What does that look like in practice?
While interest in the technology shows no signs of cooling, companies are shifting their strategic priorities for investing in and deploying it. Here are the areas we predict data and AI leaders will focus on in 2025:
Enterprise AI strategies will center on post-training and specialized AI agents
Companies will evolve how they navigate scaling laws as they shift their focus from pre-training and bigger models to post-training techniques. We’re already seeing companies build agentic AI agent systems, composed of multiple models, techniques and tools that work together to improve efficiency and outputs.
Companies will leverage agentic workflows at inference to evaluate AI systems for specialized tasks, such as debugging and improving quality over time with fewer resources and data.
“Investing in AI agents now will help organizations take a commanding lead in their respective markets as the technology grows more powerful. But few have the proper building blocks in place. AI agents require a unified foundation, free from data silos and legacy architectures.”— Dael Williamson, EMEA CTO at Databricks
Infrastructure will be the biggest AI investment area as companies race to AI agents
The Economist Impact revealed that only 22% of organizations believe their current architecture can support AI workloads without modifications. We expect to see the most resources invested in this area of enterprise data infrastructure in the coming year.
In Agentic AI Systems, agents must be able to work outside the boundaries of proprietary IT environments and interact with many data sources, LLMs and other components to deliver accurate and reliable outputs. Enterprises will need an end-to-end data platform – an AI database – to support the governance, regulation, training and evaluation required to get AI initiatives into production.
“A successful AI strategy starts with a solid infrastructure. Addressing fundamental components like data unification and governance through one underlying system lets organizations focus their attention on getting use cases into the real-world, where they can actually drive value for the business.”— Robin Sutara, Field CDO at Databricks
Companies will use their “data advantage” to gain market share
In 2024, the discourse around enterprise AI centered around internal applications that can boost employee productivity and efficiency. But domain-specific knowledge - or data intelligence - emerges as the new focus as enterprises put customer-facing applications into production. This means that companies will race to identify use cases aligned to the areas where they have a data advantage.
This is one reason why customer service is such a popular starting point. Businesses often have large amounts of data on their own clients, and can use that to power AI systems that improve the support they provide. Details on each individual’s past interactions can help personalize future experiences with the company.
But organizations can go even deeper. Manufacturers can use data assets stemming from digital manufacturing equipment to optimize the health of their machines. Life sciences companies can use their decades of experience in drug discovery to help train AI models that enable them to discover future treatments more quickly. Financial services companies can build specialized models that help clients take advantage of their deep subject matter expertise to improve their own investment portfolios.
“Companies can realize huge efficiency gains by automating basic tasks and generating data intelligence on command. But that’s just the beginning: business leaders will also use AI to unlock new growth areas, improve customer service, and ultimately give them a competitive advantage over rivals.”— Arsalan Tavakoli, SVP of Field Engineering
Governance will dominate C-suite conversations
The conversation on AI governance has so far centered on security and regulation.
Executives are now recognizing the relationship between data governance and AI accuracy and reliability. A holistic approach to governance aims to ensure responsible AI development, deployment, and usage while mitigating risks and supporting regulatory compliance.
Many companies have already taken the initial step of unifying metadata for their data and AI assets in one location to eliminate redundancies and improve data integrity. As enterprises deploy more AI use cases, this will serve as a critical foundation. Governing the two together ensures that AI models are generating outputs and taking action based on high-quality data sets. This improves the overall performance of the AI system, while also reducing the operational costs involved with building and maintaining it.
“As more businesses embrace data intelligence, leaders need to think critically about how to balance widespread access with privacy, security and cost concerns. The right end-to-end governance framework will allow companies to more easily monitor access, usage and risk, and uncover ways to improve efficiency and cut costs, giving enterprises the confidence to invest even more in their AI strategies.”— Trâm Phi, General Counsel
Upskilling will focus on boosting AI adoption
The human-in-the-loop approach to AI projects will be required for many years to come. The past two years have framed AI upskilling as needing to understand how these systems work and prompt engineering. But we’ve just scratched the surface of how today’s models can be applied, and the real hurdle to unlocking new applications is around human behaviors. That’s why organizations will turn their attention to driving human adoption - through refined hiring practices, home-grown internal AI applications, and more specialized use case training.
“In the world we’re operating in now, mindset matters more than skillset. Technology is evolving rapidly, so we need to look for people with an open, creative, growth mindset and a passion for learning and trying new things.”— Amy Reichanadter, Chief People Officer
What’s next in data + AI
2025 promises to be a pivotal year, one in which both AI and the data, infrastructure and governance surrounding it, become even more of a focus area for leaders.
To hear from 1k+ data and AI leaders about the challenges and opportunities of enterprise data management and AI adoption in 2025, check out the Economist Impact report: Unlocking Enterprise AI
Related: What the world’s largest and leading companies are using for AI tooling, top use cases by industry, and more in the State of Data + AI.