AI transformed how software is written in 2025. Data preparation is next: data analysts at companies like Amgen are already turning to AI-assisted tools instead of manually writing every SQL query. And while LLMs like ChatGPT can help generate SQL, agentic features built into data prep products are yielding even better results.   

By the end of 2026, working without AI will feel unthinkable. So we’re publishing a short series on what we’ve learned building an intelligent, agentic data prep and analysis platform and helping companies modernize from tools like Alteryx. 

Distinct AI use cases for data

It’s daunting to pick a product in a market where numerous vendors are vying for your attention, with each promising the world with AI. With increasing capability overlap between different tool categories, even starting research can feel like a challenge. 

Let’s break down AI products for data into three categories that map to the different goals different teams are working towards in the data lifecycle.

AI for core data engineering

Central IT, data platform, and data engineering teams are exploring AI tools and features to build and maintain shared, governed datasets. This aligns with the core focus of these teams, i.e. designing, building, and maintaining the systems that make data reliable, accessible, and usable for others at the organization. They’re focused on making sure everyone, from data scientists to line-of-business leaders, have what they need to perform their work, supported by a system that maintains the organization’s governance standards.  

The AI interfaces these teams are leveraging typically look like IDE coding agents (e.g., code copilots), and notebook-native assistants. 

Worth noting, these tools are not designed for day-to-day business analysis or ad-hoc prep.  

AI for business data prep & analysis

Data analysts and business data users (sitting on either analytics teams or line-of-business teams) have adopted agentic AI platforms to drive insights for business problems, leveraging data from many different data sources. These users are focused on bringing business context to the data in order to drive reporting and find actionable insights. They would like to move faster and self-serve more data preparation and analysis, removing blockers in the form of request backlogs. 

The interfaces these teams are using include visual drag-and-drop pipeline tools (sometimes referred to as visual workflows), as well as documents and SQL code. Agentic features are now being used to power all of these interfaces (something we’ll dig into in a later post in this series). 

This is the area where Prophecy is focused, replacing existing tools like Alteryx, SAS EG, and BI tool-driven data prep. 

AI for BI Querying & Decisioning (last mile)

Business consumers and executives want to be able to ask questions in natural language and get charts and graphs back from the data. Essentially, users want to cut down on busy work tasks and get to the desired outcome, faster. 

Ai interfaces in this space typically look like text/chat over a semantic model that auto-generates visuals. 

BI platforms like Tableau and Sigma are now offering AI-enhanced features to tackle these needs. Of note, while these features are becoming increasingly capable, they are still not designed to handle data prep. Prep-focused tools can better accommodate important considerations like governance that these tools are not designed to account for. 

Looking ahead to agentic AI capabilities

Changing platforms to unlock the capabilities of these AI-powered tools is still a major lift and data projects are inherently complicated. If your team is considering this, you must expect a step-change in capabilities that warrants the move. 

In the next post in this series we’ll dive into the specialized agentic capabilities that can unlock success and take a look at expected outcomes.