Domain Integration 

How AI will influence data, compute, and storage requirements will be very domain-dependent. Although AI will affect everything, it will not affect everything at the same rate. Today, software development is on the leading edge. Based on past experience, we can expect utilities to be on the trailing edge. As a result, evolution may come in fits and starts, especially because of the increased pressure on domain integration.

In the past, we tended to have domain-specific networks of communication, processing, and storage tied together by intermediaries. The intermediaries (domain connectors)  were combinations of manual staff and public general-purpose resources. As AI and Intelligent Agent systems develop, there will be a need for more and tighter integration.

An example of impending domain integration involves factory automation. In the past, factory automation systems were isolated networks of communication, processing, and storage. They were generally air-gapped from other systems. That is, there was no physical connection with systems outside the factory, etc. This was done in part because of cybersecurity concerns. If you obtained the factory robot programming, you could replicate the product and sell it more cheaply. As we move forward with AI adaptation, it is going to be more important to be efficient than to maintain air gaps.

There are going to be networks of intelligent agents that track orders, sales, inventory status, etc. That do conventional market surveillance. Recently, consumer sentiment has been tracked by measuring the proportion of Tweets expressing positive emotions vs. negative emotions. More tracking systems will be developed with new innovative tracking algorithms. Depending on the products or services, these systems can be distributed all over the planet. In addition, the systems that track how successfully advertising messages have been embedded in AI training data (see above) will predict future demand for products and services.

There will also be Intelligent Agents that track inputs, supplies, and other portions of the value chain and transportation systems. 

The resulting ‘flock’ of cooperating agents will interface with factory automation systems. Together, these Integrated Agents will make optimization decisions in real time. Thus, giving new meaning to ‘just in time’.

AI Evolution Driving Further Change 

The change scenarios above are just the beginning. We can only see so far into how AI evolution will impact networks of communication, processing and storage. But we can see that further change beyond what is described is likely. It is like standing at the mouth of the first mine where the first steam engine was deployed to pump water out of the mine, and trying to force all the changes from the Industrial Revolution. Thus, it is reasonable to expect more changes, although we can’t force what they will be.

Although we can’t clearly identify the changes that will follow these described here, we can prepare for them. The best way to prepare for this ongoing stream of change is to design architectures such that they can easily be disassembled and reconfigured as needed.

Executive Action Planning 

If you are an executive in an organization today, then you are facing these challenges. Here are some suggestions for how you can guide your team through a proper response.

First, ask them to carefully monitor what data they are collecting and what algorithms they direct their intelligent agents to use to analyze it. 

Then, make sure that the team has an ongoing process for monitoring the progress of the AI evolution. Ask the team for regularly scheduled reviews with you. This is both to underline how important it is and to keep you up to date and well informed.

Ask your team to explain how each new architecture they develop has been designed to be able to transition to the next, as yet not fully understood, architecture - easily and efficiently. Ask them to break the problem down into domains. Track changes in each domain of interest. Finally, integrate the domain views into a comprehensive solution using Intelligent Agents, conventional code, storage, processing, communications, and manual efforts to create the best architecture for today and the near future.

Help them understand that they are in an ongoing process of change. Change is difficult and anxiety-provoking. Tell your team that you know that. Just saying it out loud will give them support. More may be required. Your team may be concerned that their work will result in losing their jobs. Provisions for such an outcome may be necessary.