Dataiku Report Predicts Key AI Trends for 2025

As we step into 2025, GenAI continues to establish itself as a transformative force, redefining industries and pushing the boundaries of innovation. This year promises to be a milestone in AI advancements, with GenAI leading the charge. 

Data science platform maker Dataiku has released its annual trends report, “5 GenAI Trends for 2025: Beware the AI Agent Reckoning.” The report identifies five key trends expected to significantly impact enterprise AI. 

There is no denying that we are in the midst of a GenAI revolution. Dataiku report found that 81% of its C-suite respondents have a moderate to high level of trust in the usage of AI. Two-thirds of data, analytics, and IT leaders have invested over $1M in the last 12 months. According to Dataiku, GenAI’s current achievements are just the beginning, with even more transformative developments expected. 

Trend 1: GenAI Will Be Commoditized

A major trend for the coming year is the widespread availability of GenAI tools and AI agents. This will significantly lower barriers to entry for basic AI capabilities, leveling the playing field for undifferentiated solutions.

The report warns that “organizations that rely solely on off-the-shelf GenAI and AI agent tools risk falling into the “commodity trap,” where technology no longer provides a competitive edge, yet significant expenditure continues.”

To build differentiated AI, Dataiku recommends focusing on innovative, ROI-driven applications, developing AI agents for real-time insights, and investing in the right tooling and software.

“We’ve witnessed the rapid evolution of GenAI in an extremely tight timeframe, which has raised the stakes for companies to sharpen their AI strategies,” said Florian Douetteau, Dataiku co-founder and CEO.

“AI agents are not on their way, they are already here redefining the rules of business — and that is only going to accelerate. Company leaders have no choice but to move decisively to avoid falling into the commodity AI trap, as competitors are poised to turn their AI advantage into meaningful differentiation, business transformation, and market domination.”

Trend 2: Quantifying ROI Under Pressure

Dataiku highlights that while 72% of data, analytics, and IT executives report positive ROI for their GenAI projects, a mix of qualitative and quantitative methods are used to measure their success. Challenges such as isolating GenAI’s impact and lacking clear metrics complicate this process.

With the rise of AI agents, the task of measuring GenAI’s ROI will become even more complex. The report reveals that 85% of leaders face pressure from their C-suite to define clear outcomes for their AI investments. To stay ahead of the drive, Dataiku suggests including concrete financial and operation outcomes in the ROI, including cost savings, productivity gains, and time-to-market acceleration. 

Trend 3: The LLM Mess

Dataiku highlights that the influx of AI agents, powered by large language models (LLMs) , is creating a chaotic environment in enterprise settings.  With 73% of organizations adopting a hybrid multi-LLM approach, the challenge lies in managing these diverse models and their unstructured workflows effectively. This complexity can lead to operational confusion if not carefully managed.

The top reasons for taking a multi-LLM strategy include wanting to improve accuracy and reliability LLM diversification for flexibility and being able to avoid vendor lock-in. Notably, 79% of data, analytics, and IT executives are using a secure gateway for LLM access, with 40% using it for all LLM access and 39% for specific LLM use cases.

To navigate this landscape, Dataiku emphasizes the need for a clear strategy and governance framework for managing multiple LLMs and AI agents. By integrating new models and agents, businesses can future-proof their AI strategies and ensure long-term success. 

Trend 4: Governance as a Priority

GenAI’s rapid rise brings significant governance concerns, as noted by Dataiku. Inadequate governance frameworks expose organizations to risks such as privacy breaches, biases, and compliance issues. 

Dataiku highlights that 75% of C-suite executives worry about data privacy breaches, but the majority trust GenAI and will continue to invest in it. To balance responsible GenAI and innovation, the executives plan on establishing clear guidelines and frameworks for GenAI use, encouraging a culture of transparency and accountability in AI development, and prioritizing user education and awareness for responsible AI practices. 

Some organizations have stuck to only using one LLM and the primary reason for that has been the security and compliance issues that could potentially emerge if additional LLMs are introduced. This reveals that while the overall confidence in GenAI is solid, there are lingering concerns about its impact on data privacy and security. 

Trend 5: Upskilling the Workforce

Workers who do not resist change and learn to work with AI are best positioned to thrive and succeed in an AI-driven future. Dataiku shared that 93% of C-Suite executives believe it’s critical for employees to work with and generate insights from data. 

(Marko Aliaksandr/Shutterstock)

Organizations should not simply automate tasks; they should cultivate a culture of continuous learning that leverages the synergy between humans and AI.

However, the Dataiku report indicates that while C-level executives are focusing on operational efficiency, there’s a potential underestimation of AI’s broader transformative potential, which includes groundbreaking innovation. Change must be driven from the top to ensure success in this new era.

“The playbook for success is still being written, but best practices are already emerging to mitigate the biggest risks that AI can present,” shared Douetteau. “CEOs have a responsibility to their companies — and themselves — to get off the sidelines and make it happen. And if they don’t, they might just make history as one of the first to lose their job to AI.” 

Related Items 

AI and Collaboration at the Forefront in Domo’s Latest Report

The Top 2025 Generative AI Predictions: Part 1

Why the Future of Retail is AI Everywhere