Is Your Data AI Ready?
By Betsy Burton
Is Your Data AI Ready?
It is now nearly impossible to find a leading software offering that does not, in some way, depend upon or leverage AI to achieve its results more effectively or efficiently.
And now, organizations are going to be able to use large language models like Google LaMDA to create customized derivative models that support your business or industry.
This means, it is quickly becoming a critical issue that you are ensuring your data and content is ready for AI-enabled enterprise applications, and new AI-native applications and tools.
The question is are you ready?
It’s About Quality
I have been working with end-user organizations on data management, stewardship and governance for years. One of the biggest issues organizations continue to deal with are data integrity and quality issues; large retailers, manufacturing organizations, insurance providers and government agencies are all still plagued with quality and integrity issues, particularly with respect to historical and dynamic data.
AI by itself does not solve content integrity and quality issues. It can help verify data. But if your organization doesn’t directly deal with it and content quality, AI systems will be generating answers and advice, and learning and creating new models on bad data.
It is worse than just “garbage in, results in garbage out.” It’s more like, garbage in, result in more garbage created, learned and then generating more garbage to be learned and generated again and again.
Whatever data you use to teach and input into AI systems must be good data, or you risk creating exponentially more issues.
It’s About Richness
In addition to quality, AI systems need to learn based on a rich set of data.
If you teach a person by only reading children’s books about farms, their knowledge base is going to be really limited. The same can be said with regard to teaching and supporting AI systems; the better and the richer the information, the more valuable the responses, advice and analysis will be.
This means, to get prepared for supporting AI applications, organizations need to ensure that they are collecting a complete and rich set of information (data and content).
If you are a retailer, what types of customer information should you be collecting to better understand your customer’s context, connections, and preferences? If you are a financial investment company, are you collecting a rich set of data and content about companies and funds that you would be using with your AI systems?
It’s About Format
Last, but not least, we are seeing new database formats emerge, such as vector databases that are designed to effectively support AI systems.
Vector databases are databases in which data is stored as arrays of numbers clustered together by similarity. Vector databases can be queried with low latency, which is ideal for AI applications.
There are several relational databases providers that support this functionality, such as Postgres with PG vector and Redis. In addition, there is a market of native vector database providers emerging, including Pinecone and Weaviate.
Even if you are a conservative organization, even your traditional enterprise applications are, at least, becoming AI enabled over the next few years. In addition, your business will increasingly need to adopt AI applications to remain competitive.
You must take your data and content seriously.
To realize value from these applications, your information must be of quality and complete. To reduce risks associated with generating more bad information, your information must be of quality and complete.
Does the evolution of artificial intelligence (AI) into the workplace mean a job desert or a gold rush? The answer, in my view, is neither of these extremes. But it will absolutely change the workplace landscape and we must work on understanding and planning for these changes.
During this webinar, we will be exploring the potential impact artificial intelligence will have on different jobs and on the workforce, in general. In addition, we will be introducing Aragon Research’s new AI Technology Arc.
- What jobs will be the most impacted by AI technologies?
- How will AI change the workforce?
- What technologies do leaders need to track over the next 3 years?