Weekly API Evangelist Governance (Guidance)
As I work to help lay an opinionated foundation for what Naftiko produces, I wanted to ground my narrative surrounding automation. Automation matters in this moment, but I am not sold that everyone has the same intent when they are talking about automation, let alone selling products and services or crafting open-source solutions that deliver automation.
Definition of Automation
As with most of my work, let’s start with a basic set of definitions from the dictionary:
The technique of making an apparatus, a process, or a system operate automatically.
The state of being operated automatically
Automatically controlled operation of an apparatus, process, or system by mechanical or electronic devices that take the place of human labor.
When you ask the the artificial intelligence (Gemini) about what automation is, you get this pretty polished answer.
Automation is the use of technology to perform tasks with minimal human intervention, replacing manual efforts with systems that can follow pre-determined decisions and actions.
I feel like this is a balanced definition that suits what I am looking to produce as part of what Naftiko is attemptig to do in a very crowded and noisy market. It may not be flashy, but it is a pragmatic definition, which helps us ground the conversation around automation in what we need to accomplish in the real world.
Industrial Automation
It feel likes it make sense to take a quick look at the types of automation that are used to automate factories and other industrial practices, before we move too quickly into AI, and other forms of automation, helping us get out bearings when it comes to what many people will consider to be automation.
Fixed Automation - Perform one specific, high-volume, repeatable sequence of operations with minimal variation
Programmable Automation - Sequence of operations controlled by software changing without rebuilding hardware.
Flexible Automation - Switching between multiple products or variants with little to no manual changeover required.
Integrated Automation - End-to-end production approach which is unified into one coordinated, data-driven system.
I am more interesting in online automation but I also prefer grounding my perspective in the physical world. As you can see software does shape physical industrial automation, but I think there is also a lot we can learn from industrial automation when it comes to automating purely digital business processes.
Types of Automation
Now we get to what most people who are subscribed to this newsletter will consider to be the type of automation on the table. It is tough to narrow down the list of how you can automated as there are so many names and variations applied, but here is my current attempt at what types of automation matter right now.
Event-Driven - Events occurring across a system.
Workflow - Repeatable deterministic patterns .
Process - Actions organized in a certain order.
Pipelines - Continuous integration or deployment.
Scheduling - Occurring based upon a time.
Orchestration - Across multiple systems.
Scaling - Based upon usage and demand.
Circuit Breaker - Based upon system failures.
Testing - Running a software test against system.
Granted, there are many other ways to express automation, but I wanted to understand the foundation of automation as part of market research across API consumption and integration solutions available today. I feel this list captures the most common intent behind the word automation.
Agentic Automation
Now we get to the reason I set out to write this narrative. To understand what we mean when we say agentic automation. Artificial Intelligence (Gemini) tells me that agentic automation describes systems, particularly artificial intelligence, that possess agency, meaning they can act autonomously and proactively to achieve a specific goal with minimal human intervention—lets break down the most meaningful words used in there.
Agency - Action or intervention, especially such as to produce a particular effect.
Autonomous - Govern itself or control its own affairs.
Adaptable - Change or adjust to new conditions or situations.
Proactive - Causing something to happen rather than responding to it after it has happened.
Goals - The object of a person's ambition or effort; an aim or desired result.
That is a lot to bestow upon any automation, and I am not sure that is what everyone means when they say agentic. I will make it the bar for how I approach the word. There are many ways to automate, and there are many reasons I will need to automate, and I am not sure that I will need all of those factors for every one of my automations. But, it will become the bar for how I evaluate and consider when people use the phrase agentic automation.
Benefits of Automation
After taking another pass through the market research I am doing across 10 segments for Naftiko, I wanted to ask myself why people are automating. What are the words that services and tooling makers in the automation space were using, but also what is the intent from companies across different industries, based upon their storytelling, job postings, and other signals.
Increased Productivity
Reduced Errors
Improved Efficiency
Cost Savings
Customer Experience
Automation is our edge these days. It is how we get ahead. It is how we remain competitive. It is how we attract and maintain customers. I am focused on the why of automation first, and the how and what after that, I think if I can stay focused on the why of automation I can better speak to the real world situations where people are needing to automate—not the imaginary.
Human in the Loop
Like automation, there are different interpretations of what human in the loop means when it comes to automation. As with the types of automation I wanted to distill down into a list that were the most common, while navigating the many different ways in which humans are included directly or indirectly into common automation online and offline.
Training - Curation, labeling, prompting, and minimizing ambiguity.
Configuration - Set parameters and policies for automation.
Copilot - System proposes and the human completes each step.
Gatekeeping - Review and authorization before some actions.
Escalation - Takes over when there are problem or edge cases.
Evaluation - Checks quality of output and applies scoring.
Supervision - Monitoring dashboard and regularly intervenes.
Governance - Sets policy, defines risks, and conducts audits.
Routing - Sets the path to address conflicts in the system.
Incidents - Incident analysis and performing corrective actions.
Consent - Obtaining end users consent before taking any action.
Sustaining - Calibrating configuration and prompts over time.
This can barely touch on all of the ways in which humans are involved in automation. It doesn’t speak to the architects, developers, and other stakeholders involved in bringing automation to life. It doesn’t adequately cover the ways in which chat applications like Slack and Discord are being used to automate with humans at the center. This list is just mean to provide a starting point to think about how we talk about the human and machine overlap.
Automation Toolbox
The question I am trying to answer here is—how do we build an automation toolbox that will address what we need today as well as what we need tomorrow. I am looking to deliver integration and automation capabilities to businesses across any industry. I am looking provide solutions to folks approaching agentic automation, but also those who are strengthening their event-driven and orchestrated automation across systems. I want to deliberately use a vocabulary that speaks to the intention of legacy automation and modern automation, as well as moving between the two. This is what an automation toolbox should provide.
As part of the market research that produced this narrative I will be looking at our legacy of automation again. I am an expert in API integrations and the automation surrounding that, but I am not an expert in extract, transform, and load (ETL), business process management (BPM), and other automation disciplines. Let alone the standards and tooling that they use to get the job done. I know from experience that there is a lot we can learn about the future from studying the past, and I am confident that Nafitko will be able to strike a balance between yesterday, today, and tomorrow when it comes to our approach to the automation of business capabilities.
”The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.” — Bill Gates