AI EDUCATION: What Is a Systems Integrator?

This week we were inspired by defense, national security and foreign policy blog War on the Rocks to talk about the role of systems integrators and how artificial intelligence might change what they do. 

Simply put, a systems integrator is a person, group or technology that connects disparate pieces of technology into a cohesive—and coherent—whole. In the War on the Rocks blog, the authors discuss that in defense, there may already be too much system integration in artificial intelligence, hampering progressive development of “exquisite, unique” systems like stealth aircraft and advanced warships. This can occur because of the limited market for the hardware being developed—the U.S. Department of Defense and its allies. 

Most of the ink I’ve personally spilled on systems integration has been about putting together the ever-expanding technology stacks of financial advisors in a coherent manner. Advisors, due to their heavily regulated industry and the technological inertia that comes with an aging workforce, often depend on decades-old software and hardware to perform some of the core functions of their business, but also look to more cutting-edge client-facing solutions to keep up appearances. The hodgepodge of different levels of technology makes systems integration a challenge. 

What Is a Systems Integrator 

When the term “systems integrator” is used in the wild, people are usually discussing a person or a business (historically firms like IBM, Accenture or DXC Technology) focusing on making an entity’s different technologies work together.  

Like the financial industry, in the defense industry, different systems—like the ones that run huge warships like aircraft carriers and nuclear submarines, sophisticated aircraft like stealth bombers and drones, and coordinate the efforts of soldiers, vehicles and logistical support on the ground—were designed in so-called silos, separate from other systems. 

Also like finance, the defense industry still uses processes and equipment that is decades-old, and there is need for someone or something to step in and make sure that the new processes and equipment are compatible with all that legacy stuff to the extent that it can all be used in concert. Hence, the systems integrator. 

Now and Then 

These days, the integrator role is also filled to some extent with existing software and proprietary products—for example, in the financial world, there are different companies claiming to provide the so-called “connective tissue” between disparate pieces of software. 

Systems integrators themselves have been around for decades. In our core industry, media and publishing, from the very early days of computers we dealt with one side of media computing that was excellent and reliable at producing work for print, while another, barely compatible side of computing handled digital and most broadcast work—requiring something—in those early days often a someone who worked manually—to act as a go-between. In a large business, integrating new technology can be a process that takes years to accomplish, and can be so complex and resource intensive that, in the past, many businesses have aborted integrations in lieu of keeping manual processes. 

Today’s systems integrators are generally selling a product, the platform that they purport will make a business’s software work together optimally—disguised as what they claim to be the service of making a business’s software work optimally. In other words, a lot of systems integrators are really distributors, salespeople in engineers’ clothing. The products, until recently, were primarily enterprise service bus solutions (ESBs) or application programming interfaces (APIs) 

What Does All This Have To Do With AI? 

Businesses should consider two primary use cases for a systems integrator: implementing new technology within an existing system, or migrating from one system to another. Both of these use cases bring AI into the integration discussion, as the person or product purporting to optimize a business’s system during implementation or migration will likely need to be versed in artificial intelligence. Systems integrators will need to be able to build and augment the data infrastructure needed to train and implement artificial intelligence, especially generative AI. 

Also, a new type of systems integration, specializing in building connections between different applications of artificial intelligence, has emerged: artificial intelligence systems integration. Artificial intelligence itself is the primary tool of AI systems integration. It is this type of systems integration that links speech recognition to natural language processing and generative artificial intelligence to make possible conversational, speech-oriented chatbots and assistants like Apple’s Siri and Amazon’s Alexa. 

On a smaller scale, artificial intelligence has proven to be useful in weaving together complex systems, understanding hierarchies and assigning priorities—the kind of knowledge work that usually has fallen to human systems integrators. Generative AI can also step in and write a lot of the new code necessary to create bespoke integrations, speeding up the work and cutting the cost of piecing together a business’s technology. 

AI Will Impact Systems Integration From At Least Two Sides 

On one hand, the rollout of artificial intelligence across the economy, especially generative AI, is creating a lot of new and very complex work for systems integrators of all stripes. The role of a systems integrator is growing as more new technology emerges. 

On the other hand, AI itself is becoming a systems integrator and helping to reduce the overall workload. As new technology continues to proliferate, it remains to be seen whether AI demands more human systems integrators, or serves more to replace those jobs.