History Repeats Itself: What AI Practitioners can learn from the SQL Story

Every other day, we hear about a new advancement in AI, each one more “wow” than the previous. And yet the (short) history of enterprise AI implementations tell a different side of the story—most of them have not met the desired expectations.

So, what gives?

To answer this, we need to cut back to 1968, when a British computer scientist named Edgar Codd (whose photo is featured above) joined IBM and came up with a path-breaking idea called “A Relational Model of Data for Large Shared Data Banks.”

Codd envisioned a database that organized information into tables that could be linked based on common characteristics. This enabled users to retrieve relevant information quickly and easily, providing valuable business insights to make decisions and identify opportunities.

That’s how SQL was born.

And Codd’s approach offers a great learning opportunity for AI practitioners.

The main reason why AI implementations don’t succeed is because organizations are unable to effectively leverage their data, most of which is unstructured and contains valuable insights—but is the most neglected. ?While 90% of data is unstructured, organizations only allocate 40% of their resources to extracting its value.[1]

Taking a leaf out of Codd’s approach, we proceeded to classify ALL enterprise information and knowledge and found that they fall rather neatly into three buckets:

1.?Data, which includes all structured and unstructured information

2. Code, which encompasses the entire Software Development Life Cycle, and

3. Experiences, which include audio, video, and text-based knowledge.

Next, we applied Codd’s structured querying to provide the much-needed consistency and reliability to prompt engineering, replicating the level of systematization that SQL brought to database queries.

The result is what we call the Activation AI approach.

Activation AI aims to help enterprises take control of all their structured and unstructured data, business information, and knowledge repositories, analyze them using the most relevant AI models, and derive actionable business insights—without breaking the bank.

In 2002, Forbes listed Codd’s relational model as one of the most important innovations of the previous 85 years. One might add, applying the same ideas could well be the most important innovation for the future of enterprise AI as well.

Today, 200+ projects later, the Activation AI approach seems to be working. Some say it’s path breaking but I invite you to check it out for yourself.

Let’s connect and I’ll be happy to demonstrate how it could work for you.

#ActivationAI #AIthatWorks #EnterpriseAI

Image credit: Fair use, https://en.wikipedia.org/w/index.php?curid=2151320

References:

1.According to IDC, Box Inc.