Looking Ahead: AI Hurdles IT Leaders Need to Overcome in 2025

Dataiku Product, Scaling AI, Featured Renata Halim

In May 2024, Dataiku and Databricks surveyed 400 senior AI professionals from large companies worldwide to track shifts in the data and AI (including Generative AI) landscape since last year’s survey analysis. IT leaders continue to play a critical role in scaling AI, and this year’s insights highlight new challenges and priorities within IT and information security departments.

While IT leaders remain a smaller segment of respondents (n=80), their perspectives are essential for understanding the evolving obstacles in scaling data and AI projects. So, what’s changed since last year? Which barriers have grown more pressing as companies strive to realize AI’s full potential? Here’s what we found.

→ Read Now: AI, Today: A Survey Report of 400 Senior AI Professionals

1. 50% of IT Leaders Lack Visibility and Control of Data, Analytics, and AI Projects Across the Organization

In 2024, half of surveyed IT leaders reported significant challenges with visibility and control over data, analytics, and AI initiatives within their organizations — an increase of 13% from the previous year. This rise underscores the ongoing struggles IT departments face in managing complex data environments effectively.

⚡For further insights, check out our blog on the challenges that keep IT leaders up at night.

2. 35% of IT Leaders View Cost as a Barrier to Preventing Organizations From Delivering More Value From Data, Analytics, and AI

Cost concerns have become increasingly prominent, with 35% of IT leaders now identifying it as a barrier to scaling analytics and AI —  a significant jump from 18% in 2023. While cost has seen the most notable change amongst this audience as a value-preventing barrier, it hasn't shifted significantly enough to impact overall analytics and AI trends. Instead, it points to the growing challenge of balancing innovation with budget constraints as GenAI applications continue to expand.

As IT leaders work to manage the financial impact of GenAI, many are turning to tools that can help them oversee costs and mitigate risks. Dataiku helps address these challenges by:

  • Anticipating financial impact and managing operational costs: As more large language model (LLM) applications are deployed across enterprises, monitoring performance and costs becomes crucial for IT teams aiming to maximize ROI. Dataiku’s LLM Cost Guard offers a way to track and control expenses by application, service, user, or project, helping organizations avoid budget overruns and make the most of their GenAI investments.
  • Embedding governance to prevent future financial and compliance risks: The adoption of GenAI introduces new layers of complexity, from managing LLM costs to ensuring regulatory compliance with frameworks like the EU AI Act. Without a governance framework or usage controls, organizations face heightened operational risks that can hinder the modernization of analytics and AI. Dataiku Govern addresses these challenges by embedding compliance, documentation, and risk management into AI workflows, enabling organizations to scale GenAI responsibly while minimizing financial and regulatory risks.

⚡ By proactively managing cost and governance challenges, IT leaders can drive sustainable AI growth, balancing innovation with budget and compliance priorities. Explore more in our survey report for IT leaders, “A CIO’s Guide to Modern Analytics”.

3. 49% of IT Leaders Lack Quality Data or the Ability to Easily Grant Access to the Right Data

While the percentage of IT leaders that grapple with data quality and access challenges has dropped from 55% to 49% over the past year, nearly half still face significant barriers in managing and accessing high-quality data across departments. While this is a positive sign, challenges persist, particularly in supporting analytics, machine learning, and GenAIinitiatives.

⚡Check out this flipbook to learn five essential steps to improve data quality, featuring actionable steps like eliminating data silos and enhancing data governance to ensure accuracy and consistency across data assets.

4. 49% of IT Leaders Struggle to Operationalize and Iterate on Analytics & AI Projects

Last year, 61% of IT leaders faced challenges in operationalizing and iterating on analytics and AI projects. At the time, we highlighted the importance of alignment between IT and business teams as critical to ensuring AI solutions meet specific business needs. A lack of alignment often led to low adoption rates and diminished trust in AI-driven insights.

This year, that number has dropped to 49%, showing some progress in establishing efficient workflows and fostering better communication between teams. However, nearly half of IT leaders continue to encounter obstacles — especially in maintaining data quality throughout the process. As the saying goes, “garbage in, garbage out” — without quality data, even the most sophisticated AI models may produce unreliable results.

As we emphasized last year, successful AI operationalization requires a seamless process for integrating, testing, deploying, and monitoring models. Inconsistent data quality or deployment practices can lead to performance degradation once models reach production. Aligning IT and data teams, particularly when they share tools and objectives, can ease this transition and ensure AI models are adaptable to business needs.

To address these challenges, IT leaders can foster cross-functional collaboration and leverage Dataiku, which offers embedded, as-you-go data quality infrastructure. This infrastructure allows everyone within an organization to play a role in data quality, operationalizing it across the analytics and AI lifecycle. This approach builds data literacy by providing teams with a shared understanding of data quality at every stage, ensuring data used in analytics, AI, or GenAI is accurate and trusted. It also grants greater control to quickly identify and proactively resolve data quality issues.

⚡Learn effective strategies for operationalizing data quality and achieving consistent AI success.

Navigating the Future of AI

As the landscape of AI continues to evolve, IT leaders must remain vigilant in addressing the persistent challenges of visibility, data quality, and cost management. The insights gathered from this year's survey illuminate the pressing obstacles that can hinder the realization of AI's full potential. However, by fostering collaboration, leveraging advanced tools, and embracing innovative practices, organizations can turn these challenges into opportunities for growth.

You May Also Like

No-Code ML and GenAI With Dataiku and Fabric

Read More

Unpacking 3 of the Biggest Controversies in AI Today

Read More

4 Strategies Every CIO Needs to Succeed With GenAI

Read More