Disruption is imminent. Is your company prepared? Your IT department? You?
What do bank tellers, assembly line workers, retail clerks, blacksmiths, and travel agents have in common? They have all been largely displaced by technology. Application developers are next. In just a few years Generative AI will produce most of the code, and IT departments full of coders will as seem as quaint as data processing centers full of punch card operators or ledger rooms full of accountants wearing green eyeshades.
Generative AI applied to application development is an imminent disruption.
It is already underway. IBM is currently running television advertisements for a product that leverages Generative AI to write code given an English language specification.
This is the continuation of a long-established pattern. Binary encoding was simplified and accelerated through assembly language. Then came imperative, procedural, and object-oriented languages, followed by no-code / low-code and graphical development platforms. Common application tasks have been abstracted and consolidated into libraries. At each step, the result has been more rapid deployment of increasingly sophisticated capabilities.
But with this application development acceleration came commoditization. Once integrated into a standard library, your genius network data transfer protocol was no longer the competitive differentiator it once was. Not only did that capability become available to everyone, they could deploy it much faster than you did. This is another long-established pattern. In the early days of computing, cutting-edge companies leveraged computers and built data centers. In time, data centers became a ubiquitous corporate infrastructure component. Now, the data center itself is being commoditized in the cloud. Analytics has followed a similar path from MIS department to data warehouse to data lake to advanced analytics to AI/ML.
Yesterday’s competitive differentiators are tomorrow’s commodities.
When I was in school, I had to code neural network training and execution algorithms by hand. It took a day to translate the equations into code and to get it all running correctly. For fun, I asked ChatGPT to “Write a C program to set up and train a 3 input, 2 hidden, 3 output node neural network.” In less than 10 seconds I had code that compiled and ran successfully. I then asked for the same thing in pascal and python. Recall from Neural Network (and ChatGPT) Magic Secrets Revealed that this code doesn’t exist anywhere explicitly within the ChatGPT neural network. No database record containing the code was retrieved. The response was generated word by word.
New application development skills will be required just to keep up.
The first step in the development process will be to express the business requirements in a way that properly prompts the AI to generate the initial code. The line between those on the business side that develop the requirements and those on the IT side that interpret them will blur and then disappear entirely into the ultimate realization of the Agile partnership. Developers will still be required for the time being to customize, validate, and harden the code, but it’s not a stretch to imagine AI not just writing the code, but also compiling, testing, deploying, and monitoring it, too.
Much has been written about competitive differentiation, but here I’ll focus on a specific perspective: What in your IT estate is unique and cannot be commoditized?
Most companies will not be differentiated by their technology or their applications or the skill of their developers. Your competitors can use the same hardware and cloud platforms. They can use the same development languages, integrated environments, tools, and libraries. They can use the same software packages and SaaS applications. They can use the same databases and analytical algorithms. But no other company has your data.
Data is the foundation of competitive differentiation.
But just having data isn’t enough. You probably recognize the aphorism, “if you can’t use the data, you might as well not have it.” I would extend that statement slightly to say that “if you can’t use the data quickly, you might as well not have it.” And to use the data quickly you must understand it.
The business must understand the data to put it to use. Developers must understand the data to accomplish the AI-generated application customization and validation. Gaps in understanding reduce velocity and flexibility, and therefore reduce potential competitive advantage.
This also means that the need to understand data is no longer limited to an information management group or Center of Excellence, or to Data Analysts or Data Stewards. Everybody must understand the data. Everybody is responsible for protecting it, securing it, and ensuring that it is accurate. Data is everybody’s business.
Everyone is now a data person.
Your company needs to begin preparing for this reality now if it hasn’t already. The alternative is falling further and further behind your competitors that are. Data focus must be real, not just lip service. It doesn’t help to pretend. If application development and delivery is still the primary focus of your IT organization, it is on a path to obsolescence.
