The Robotics / RPA / AI / ML / Data / IaaS Stack

Robotics RPA AI ML Data. IaaS stack

We’ll be covering a lot of ground in the upcoming series of posts so I wanted to lay out a bit of a map of where we will be going over time.

Like in the early days of cloud computing, there is much confusion around Robotics, RPA, AI, ML, big data, and even infrastructure. I think that the best way to think about it is hierarchically. Robotics for example requires (mostly) AI underneath it. AI depends on Machine Learning as one component. All of those sit on top of data sources which in turn sit in database and analytics systems. And that of course sits on infrastructure.

The entire stack – applications, components, and infrastructure

I’ll expand on each layer of this below in order of depth.



Robotics requires not only AI but also mobility and overall development and operational tooling. This is a VERY lightweight map – I’ll develop this more fully later. The key is just to understand the overall hierarchy. Deeper levels of each sub-map will be developed over time.

AI is one of many components within Robotics

Robotic Process Automation

Robotic Process Automation is a terribly named set of practices that does not actually use Robots. Or at least not physical ones. It is really the next generation of what we used to call BPM (Business process management) – essentially the automation of back-office systems, often by tying together legacy systems. It is based on data and automation but can also be augmented with AI, hence I’m adding it to this hierarchy.

RPA does not necessarily require AI but AI can provide additional leverage

AI Use Cases

Use Cases by Sector

AI will likely impact every sector and every size of organization

Use Cases by Function

AI will impact almost every major business function. Here are a few examples:

This is a small sampling of use cases in each functional area


Artificial Intelligence Platforms

Think of AI as yet another development environment, like database development or app development. It’s not useful until you have a function and a sector and a business problem to apply it to. Below you can see a few representative micro-applications built on top of the core services which are in turn built on the underlying frameworks.


All of these systems require data. The type and volume may vary but I’ve put some representative sizes below that seem to appear in many discussions.

There is a reason that the phrase “Data is the new oil” exists – it’s key for AI training

Databases and Analytics Systems

Cloud-hosted databases, DB-as-a-Service, and even Data warehousing as a service (DWaaS) have exploded in the last 10 years. There are now a lot of different ways to store, transfer, and access data.

Beware of data gravity. It’s easy to build data up in one spot and then it costs to get it back out!


Phew! Finally we are at the bottom of the stack. AI sits on big data and big / cheap / fast compute and storage. This is what the Infrastructure stack looks like.


Now that we have a rough framework, this should make future posts make more sense and help people see the forest for the trees. I’m sure that there are many ways this stack could be drawn, but this is my attempt to bring some hierarchy and order to it for the purpose of this series of posts.

If you think I’ve missed something or am wildly wrong, let me know in the comments! And thanks for following along on this AI series.