How would you define “AI-ready” data? Is your D&A ecosystem ready for practical AI implementation? If not, what do you think is needed to get there?
Sort By:
Oldest
Partner / Principal in Services (non-Government)6 months ago
AI readiness is a topic many organizations are grappling with. One emerging theme is using AI to define the readiness of data. While most organizations struggle with disparate data systems, the power of generative AI can help unlock value from this data while improving its maturity. The key is not to look at structured and unstructured data separately. Bringing them together opens up a new paradigm of what the definition of readiness would look like.Chief Strategy Officer6 months ago
Leaders are now recognizing that data has traditionally been siloed within departments. This is problematic as a deep learning model is only as good as the data fed into it. The realization that data isn't ready for AI could serve as a wake-up call for organizations to implement an enterprise-wide data policy.We're currently working with companies who are devoting two to three months to clean and align their data before starting on an AI project. This is a trend I've noticed changing in the last four to five months as companies seek to get more context out of large language models. If the data isn't properly aligned and cleaned, it takes a lot of time and the results may not be as accurate or useful.
Chief Data Officer in Travel and Hospitality6 months ago
The readiness of AI data is a complex issue. The first step is understanding the business process and decision we're trying to automate or augment with AI, and then determining what data is necessary to support this decision. Too often, discussions are led by technology rather than by the decision we're trying to automate.When it comes to data, especially unstructured data, finding biases and inaccuracies can be challenging. Most organizations struggle with managing structured data, let alone unstructured data. If an organization hasn't mastered their structured data, they're likely to encounter difficulties when trying to govern unstructured information.
In my opinion, most companies aren't in a good place with their unstructured data. The algorithms we're using to pull this unstructured information together are revealing gaps, potentially security gaps, in old information that's still sitting around because it hasn't been managed very well.