Query & Theory Newsletter: January 2025

With 2024 behind us and 2025 stretched out ahead, the data and data science community has been reflecting on what changed in the past year and where things are headed. There is certainly a lot of alignment around the rise of AI agents (perhaps too much at times) but there are also some interesting nuggets, specifically around interoperability of AI agents across the enterprise stack, and where to invest time developing new skills to succeed in the future.

[Community post] Aaron Levie chimes in on enterprise AI finally emerging

Aaron Levie, CEO and founder of Box, shared some thoughts on X about the state of AI in the enterprise, and his comments seemed to have really resonated. He calls out specifically that AI is starting to emerge in the enterprise and is starting to make an impact across all departments. There’s been more hype in AI in the past two years than almost any other technology in the past decade, but we’re starting to break through the noise. We can attest to this as well at Fabi.ai, where we’re seeing our platform adopted across the enterprise and truly make a difference.

Aaron also calls out that enterprises want their choice of AI and that AI solutions will have to learn how to operate before different enterprise systems in order to be effective. Only time will tell how nicely large enterprise providers will play together.

[Video] Pushing dynamic alerts and data to Slack

Customers often ask us if there’s an easy way to push reports to Slack, or better yet, create dynamic alerts based on the data. For example you may want to only send a notification if a certain threshold is met, or perhaps you want to call out a risky KPI with a specific type of emoji if the number dropped.

In this 10 minute tutorial, we show you exactly how to do this step-by-step.

[Podcast] 2025 AI and data science predictions

In a recent Super Data Science podcast episode, Jon Krohn and Sadie St. Lawrence review their predictions for 2024 and look ahead to 2025. Notably, they predict that:

• 2025 will be the year of the AI agent (isn’t an AI agent just a workflow anyways?)
• LLM providers will start to push harder on monetization and cost reduction
• “The demand for AI engineering skills will surpass traditional data science skills”

That last point is a great clue about where to invest time learning new skills, which pairs nicely with the post from Kevin Hu shared below.

[Community post] Interest in data engineering skills is sharply increasing

In a recent LinkedIn Post, Kevin Hu (co-founder of Metaplane), highlighted some key findings from the O’Reilly 2025 tech trend report. He calls out specifically that interest in data engineering skills has risen by 25% while Python has dropped by 5.3%. Successful AI in the enterprise is highly dependent on good data, which starts with strong data engineering skills. On the other hand, as much as AI is great at generating Python (trust us, we know!), learning the fundamentals remains a great way to stand out and stay ahead of the pack.