This article is built from a subscriber’s question about the differences between digital and intelligent transformations. Don’t forget to take advantage of your subscriber benefits (see more at the end of this article):
Content Built From Your Questions And Comments.
Weekly Office Hours On Mondays 6 pm PT And Fridays 9 am PT.
A 15% Discount On Courses, Reserved Spots, And Early Access.
Intelligent transformation can use digital transformation as the jumping-off point. While they are different, they shouldn’t be separate.
In my Continuous Transformation framework, intelligent transformation is just the next step. Digital transformation has activities like process discovery and mapping that intelligent transformation benefits from. Digital technology is the layer that gives businesses access to first-party data. Process discovery creates the connection between business context and data gathering. Without them, the data cannot be used to train models efficiently.
This is one boundary condition that differentiates digital from intelligent. In the digital paradigm, data is gathered for people or applications to consume. In the intelligent paradigm, models are the data consumers. Data can be gathered without context when people and apps are the end consumers. This is a critical difference and where I will start the article.
A Technical View Of Digital VS Intelligent Transformation
In the digital paradigm, people provide the context around data. We evaluate the data and determine what should happen. That’s also true when apps consume the data. A human operator is often on the other end of the app and controls the workflow.
Even in digital automation scenarios, people still provide the context. Applications are built to a specification. People provide the context for what the application should do. That context is turned into code that captures the business logic for the workflow. The code represents some of the domain knowledge required to complete the workflow.
In digital automation, data allows the application to be more responsive. Still, on the back end, business logic determines what action should be taken. Conditional statements capture the domain knowledge so the app can execute the proper steps in response to the conditions.
This gets very expensive as the number of conditional branches increases. Any workflow that takes many data points and has dozens of possible branches requires a lot of code to automate. Each layer of complexity adds cost and increases the likelihood of something breaking. There’s a complexity threshold where digital solutions are no longer feasible.
That’s where data and AI take over. Data is gathered with as much business context as possible. Models are brought in to learn the context and respond to conditions like the application would. The data used in digital applications has a very low context signal, so it takes a lot of it to reconstruct the context. Data gathered with context has a higher signal, so reconstructing the context takes less of it.