How to prepare data for whatever the future holds for AI

Is there a way to keep data constantly AI-ready—even when the next AI trend is largely unknown?

Ramon Thomas, a strategic advisor at IT services and consulting company World Wide Technology, told IT Brew that “data is the crux of all AI initiatives,” while acknowledging that many organizations have historically struggled with data quality in general.

Poor data quality can lead to a “garbage in, garbage out” situation where poorly structured, undefined data ingested by an AI model leads to poor results, according to Thomas.

According to a 2026 Cloudera report, only 7% of respondents (which included people involved in their organization’s data decisions, and whose organizations were actively exploring, testing, or running AI projects) said their organization’s data is completely ready for AI adoption. Over one-quarter (27%) of those surveyed said their data is either not very or not at all ready.

Eryn Brodsky, practice lead for servers, storage, and platforms within the hybrid infrastructure organization at CDW, said that while certain industries like finance and healthcare are ahead on the data maturity and quality curve, she estimated that the “majority of people” are unsure what the breadth of their data looks like or where it’s stored.

“Start there, because if you don’t take the time, you’re going to end up having a challenge in the future,” Brodsky said.

What makes data ready for AI? Brodsky said that “AI ready” can be defined as information that has been “properly prepared” for AI utilization through proper securing and cataloguing, adding that preparation flows in “roughly” seven different phases:

• Discovering what data an organization has and where it’s stored
• Moving the team’s focus from inventory to metadata to find where the data has been and if it’s been catalogued
• Figuring out the quality of the data and its level of cleaning
• Aligning the data’s access and privacy components, as well as security
• Preparing the data for AI ingestion, which includes more indexing
• Quality-assurance sign off from humans in the loop
• Monitoring and adjusting as needed

This tangible path can make it “a lot easier to identify what data is AI ready, what data is not, and what are the actions we can take in order to complete any gaps that we’ve identified,” Brodsky said.

Alteryx, a data analytics and automation platform, defines AI-ready data as information that is “prepared for a specific AI use case, with the context and controls the system needs to use it well.” Models must translate values in context, meaning that vague information with insufficient or wrong context can lead to poor AI results. 

Teamwork. Planning for future AI trends and models means understanding data use. Brodsky said IT pros should consider collaborating with other teams to understand how their organization as a whole is using and leveraging data.

“Those are the things we sometimes forget about,” Brodsky said. “That cross-team collaboration is the thing that can really make the difference when trying to plan for your future and how you’re using AI.”

Thomas said enterprises can consider bringing on a data curator to help identify relevant data for AI projects; alternatively, IT teams can use programming frameworks to identify the types of data needed for AI models and data explainability.

Thomas said that, in his experience, professionals should also lean into auditing data stored within an organization. Some organizations may struggle to understand data sources and how to build policies for identifying and leveraging information because “there’s just such a robust amount of information that they’re garnering there.”

“You can’t apply policy and governance to what you don’t know,” Thomas said. “Putting in tools and resources that allow you to conduct that auditing effectively is just absolutely crucial.”