47) AI vs Automation

Behind-the-scenes building Vambrace AI, a company on a mission to figure out its mission. Please pardon the stream-of-consciousness style. Subscribe to follow along or visit the site here:

(typos are to make sure you’re paying attention)

Introductory Remarks

Dear Vambracers —

In last week’s post, Lower Middle Market, I explored the company’s new market positioning as specifically serving “emerging” PE firms that acquire lower middle market companies. In short, it seems like I’ve found a real niche that aligns with my background and interests, and it’s nice to be speaking to market participants that are receptive to my pitch. Moving on!

AI vs Automation

In today’s post, I wanted to briefly discuss the difference between AI and automation. In the popular tech discourse, I think “AI” has become a catch-all term to describe really any automation-based process improvement. But there’s a real difference between AI and automation.

AI

Artificial intelligence is basically a statistical model that got really really really good at predicting the next word. From there, we realized that next-word prediction pretty much underpins everything in the digital world—because code is just words—and then eventually everything decomposes to 0s and 1s anyway, and AI is really good at predicting 0s and 1s. Then, at a certain point, prediction sort of transforms into creation and generation—and voila, we have the ability to create pretty much anything digital.

The real promise of AI stems from its ability to navigate through complex and large datasets to discern meaning and create classifications and arrive at some output. There is some degree of “human-like” reasoning there, where it can display (or simulate) thinking to help it make sense of the data.

Automation

On the other hand, “automation” refers to any general process improvements that convert manual work into automatic processes. This is more about taking information from some platform and then putting it in another, or data labeling, or statistical analysis, etc., basically any business process where there is some perfectly specific desired output.

The Difference

I’d contend that AI is a subset of automation, and that the difference really lies in the sophistication and complexity of decision-making within the context of some business process. Automations basically rely on deterministic decision-making where there isn’t any uncertainty, randomness, or reasoning required to transform some input into some desired output. Whereas the promise of AI is that it can apply non-deterministic reasoning to complex input data to arrive at some desired output. The challenge is that AI still makes many mistakes, and everyone is working on different ways of guiding AI through pragmatic reasoning processes to drive business outcomes.

The Discourse

As I mentioned earlier, AI has become a catch-all term for pretty much any process improvement—when oftentimes AI isn’t actually the animating technology for some end-use process improvement. In many cases, AI can be used to build automations more quickly than would be possible in the past—and the universe of folks who can now build automations leverage AI to do so. So there’s an influx of folks who can build automations, coupled with increasing general public awareness of AI—the revolutionary technology.

But now, as companies explore the application of AI within their companies, in many cases what they’re asking for really isn’t “AI” at all, it’s just some deterministic process improvement that AI makes easy to build. Because the truth is that there are many deterministic and manual processes that still exist within organizations that humans have historically been asked to do.

Now that we can easily and cheaply create custom software to automate deterministic processes, there’s no real reason that humans should be looking at one screen, applying some basic transformation (if any transformation at all), and then typing that information into another screen.

So really as business executives feel external pressure to “leverage AI,” in many cases—particularly in the lower middle market—it serves more as a catalyst for critical self-reflection and cajoles leaders to invest in change. And that’s the opportunity.

Looking Forward

It’s been fun to experience life on the frontlines of “AI implementation and adoption” when really in many cases I’m just helping companies build automations, and I’m using AI to build those automations quickly and cheaply. And from there I’ll be able to build more sophisticated AI-animated automations as the technology improves and appetite deepens.

Have a great week!

Sincerely,

Luke