Follow Datanami:
December 8, 2023

Is Your Data Management Strategy Ready for AI? 5 Ways to Tell

Daniel Zagales

(Lerbank-bbk22/Shutterstock)

Next-generation AI technology is to enterprises what smartphones are to teenagers: Something almost everyone wants, but not everyone is ready to use properly.

To the typical teenager (or tweenager), smartphones are a must-have commodity that signals maturity and an embrace of the future. But to own a phone responsibly, teens should demonstrate certain levels of social and technological awareness.

In a similar fashion, most enterprises today are eager to embrace modern AI technologies, such as large language models. But depending on the maturity of their data management practices, they may or may not be ready to go about implementing such technologies.

Now, I’m not here to give you advice about the fraught decision of whether to buy an iPhone 15 for your teen. But what I am qualified to do – because it’s something I advise businesses about on a regular basis – is help you decide whether your business has the level of data management maturity necessary to take full advantage of AI.

Why Now is the Moment for Modern AI

(PopTika/Shutterstock)

Unless you’ve managed to ignore tech news for the past year, you know that AI is among the hottest trends at the moment. The growing maturity of generative AI technology, in particular, has made enterprises acutely aware of just how much potential modern AI solutions bring to the table – and how much of a competitive edge they stand to gain or lose depending on how effectively they embrace AI technology.

Indeed, even if your business has not yet begun experimenting with tech like large language models, now is the time to start. The window for gaining a competitive edge with help from next-gen AI is still open, but it won’t remain that way forever, and businesses that don’t act soon will be left behind.

Is Your Data Management Strategy Ready for AI?

But again, just because you want AI doesn’t mean you’re ready to use it. Before jumping on the AI bandwagon, businesses need to assess their data management strategy and evaluate how well it positions them to deploy and use next-generation AI technology.

A full discussion of items to consider in this context is beyond the scope of this article, but here’s a look at five of the most important factors.

1. Data Quality

The old mantra about “garbage in, garbage out” is especially true in the context of AI. If you train an AI model using low-quality data, your model will make low-quality or inconsistent decisions.

via Shutterstock

That’s why being able to assess and optimize data quality is a key requirement for making use of AI. Data quality starts with defining metrics that reflect how accurate, complete and consistent your data is, then measuring those metrics on a routine (or, better, continuous) basis. In addition, you should have tools and processes in place to improve data quality by, for example, removing redundant information from data sets or excising outliers that may represent inaccurate data points.

2. Data Accessibility

You won’t get very far with AI if it’s difficult for AI algorithms and models to access your data. For that reason, you need a data management strategy that ensures data accessibility, meaning that all data owned by your business can be easily connected to or integrated with applications (including AI apps) that want to use it.

The importance of data accessibility is often obvious when you’re dealing with “ordinary” types of data, like databases. But keep in mind that data accessibility is also critical for other types of data, such as semi-structured, unstructured and “dark” data, all of which may also play a role in AI use cases.

3. Data Flexibility

Data that is available only in one form and can’t be restructured is not very useful for AI. Neither is data that you can only access on a small scale or under certain configurations.

To make the most of AI, you need data that is as flexible as possible. No matter which volume of data you are working with, how it’s structured or where it’s stored, your data management tools and processes should allow you to apply the data to any AI use case. Sometimes, doing so will require making changes like migrating your data to a new storage platform or converting it to a different format.

4. Data Governance

One of the key challenges related to modern AI is that you don’t always know what an AI model does with your data, especially if you’re working with third-party AI services.

That’s why data governance is a critical pillar of responsible AI use. Data governance allows businesses to establish rules about where and how different data assets can be used. You may have some data that is too sensitive to expose to a third-party AI service, for example. With data governance policies, you can lay out clear policies that define how AI models can use data.

4. Data Stewardship

Establishing data governance rules is one thing. Actually enforcing them is another – and that’s where data stewardship comes in.

Data stewardship allows you to implement processes that ensure that your teams follow data governance and quality rules when working with data. Proper data stewardship protects against the risks posed by AI models, among other challenges.

Making the Most of Data in the Age of AI

Data has long been a critical element in business success. But next-generation AI technologies have made it even more important for organizations to take full advantage of the data at their disposal.

However, as I am reminded on a daily basis as I work to help companies modernize their data management strategies, some businesses are more prepared to leverage AI than others, because some have more mature data management practices in place than others. Before trying to embrace AI technology only to discover your data management strategy isn’t mature enough to support it, assess how you manage your data, then identify and address any gaps before you begin implementing AI.

About the author: Daniel Zagales is the VP of Data Engineering at 66degrees, where he is responsible for defining and executing on strategy and go-to-market in the Google Cloud Data Management and data and analytics pillars; defining and shaping 66degrees methodology and best practices; providing support to key accounts through a pre-sales and delivery; and defining growth paths for 66degrees associates.

Related Items:

Data and Algorithms: The Building Blocks of Artificial Intelligence

How Hybrid AI Will Enable Far Greater AI Use

Will Mass Adoption of GenAI Lift Traditional AI?

Datanami