Unit Introduction
Welcome to the first unit of the course AI, AI agents and automation.
So far, you’ve gathered a lot of information about what AI agents are and how they work. Now, the time has come to start exploring AI agents in
Make and how you can use them in automation.
You will learn: the differences between traditional automation, AI automation, and agentic automation
how they make decisions
the strengths of each automation type No time to lose!
Types of automation
In Make, you can use AI agents to run your scenarios. When talking about AI agents in Make, you will hear the term agentic automation. It can sound very fancy, but it simply means using AI agents in your scenarios to complete tasks on their own, without needing you to guide every step.
Need a quick recap on AI, generative AI, and LLMs?
Check out AI Fundamentals and Mastering LLMs. This introduces a new kind of automation alongside traditional automation and AI automation. What are their characteristics?
TRADITIONAL AUTOMATION AI AUTOMATION AGENTIC AUTOMATION
Traditional automation is what Make is known and famous for. It’s a workflow that follows fixed rules to do simple, repeated tasks automatically.
For example, it can send a message every time someone signs up on a website, or save all the email attachments to Google Drive.
It’s like Eddie the Evidence Logger, who, whenever new evidence is discovered, enters it into the system using a fixed template. Always the same steps, no exceptions.
TRADITIONAL AUTOMATION AI AUTOMATION AGENTIC AUTOMATION AI automation is like an improved version of regular automation. It incorporates generative AI in one or more steps, using LLMs to carry out specific tasks.
For example, AI can review messages sent through your website and determine who should handle them. The scenario then sends each message to the right person or team so users get help quickly.
It’s like Ingrid the Interview Assistant, that helps Derek the Detective. She provides Derek with basic interview questions, but also suggests tailored questions based on suspects’ profiles and past interviews.
TRADITIONAL AUTOMATION AI AUTOMATION AGENTIC AUTOMATION
Agentic automation uses AI agents that make their own decisions and adapt to changing situations. Instead of being told each step, they figure out the best way to reach a goal using the information they receive.
For example, a marketing AI agent could track social media mentions, analyze sentiment, and update the marketing team with insights. It adjusts its actions as trends shift, all without needing step-by-step instructions. Or, think of a recruiting AI agent. It can automate tasks like sourcing candidates, screening resumes, engaging with applicants, and handling messages, making the hiring process faster and easier.
It’s your friend Derek the Detective. He doesn’t just ask fixed questions, he investigates every clue in different ways: as new evidence shows up, he changes his approach. He is also able to figure out the best actions on the fly. This is what makes him so good at solving cases!
To better understand the difference between traditional automation, AI automation, and agentic automation, let’s have a look at the kind of logic they use.
Logic, in this case, is how a scenario makes decisions: whether it’s following strict rules or making flexible judgments based on the situation. Make scenarios can use two types of logic. Boolean logic
Boolean logic works with true or false values, following clear, predefined rules. Every action depends on checking conditions that have a yes/no answer, there’s no room for interpretation.
For example: If stock is below 10, reorder trench coats. This means that if a specific condition is true (i.e. stock below 10), perform a specific action (reorder trench coats).
This logic is deterministic: given the same input, the scenario always behaves the same way.
When Eddie finds a new piece of evidence, he asks: “Is this evidence a fingerprint?”
If yes, he fills out the fingerprint section of the template.
If no, he skips that section and moves to the next.
Some might say that Eddie’s job is not that interesting, but it’s all elementary my friend.
Fuzzy logic
Fuzzy logic works differently. Instead of strict yes/no answers, it allows for degrees of possibility. It handles uncertainty, ambiguity, and nuance, things that are hard to express with true/false rules. It asks questions like: How likely is this email spam? Or, Does this message sound angry? This makes fuzzy logic non-deterministic: the same input might lead to different outcomes depending on the context or new data. The scenario can adjust its actions based on new facts, improving its decisions over time.
Ingrid the Interview Assistant sits across from Tyler the Thief, carefully watching his reactions. When Tyler’s tone shifts, sounding nervous, Ingrid notices and adjusts her questions to be gentler, encouraging him to open up. She’s the good cop in this case. Do not trust Ingrid, Tyler!
Which kind of logic do the three types of automation use? Click each card to find out!
Boolean logic Boolean logic and fuzzy logic
Fuzzy logic
Traditional automation uses Boolean logic because it follows fixed, clear rules that produce predictable outcomes. This works well for simple, repetitive tasks where everything can be defined as yes or no.
AI automation uses Boolean logic to handle clear, straightforward decisions with yes/no answers. At the same time it uses fuzzy logic to manage uncertainty and nuance, such as interpreting emotions or probabilities. In this case, AI enhances specific tasks within a predefined workflow by processing data, interpreting unstructured inputs like text or images, and helping make decisions such as classifying information or extracting key details.
Agentic automation relies mostly on fuzzy logic because it needs flexibility to adapt, learn from new information, and make decisions in unpredictable or changing environments where strict rules don’t apply.
The three kinds of automation have different
strengths:
traditional automation is fast and reliable for
simple tasks;
AI automation combines clear rules with some
flexibility to handle more complex situations; agentic automation offers broad autonomy to
handle changing situations.
Here’s a recap of the three types of automation, with indications of when to use each. You’ll explore them in more detail later.
Traditional automation AI automation Agentic automation
When: Use when your When: Use this for When: Use when tasks tasks follow clear, mostly rule-based are complex, simple rules and don’t tasks that need some unpredictable, and change often. smart decision making need human-like or pattern recognition. reasoning or full Why: It’s fast, reliable, independence. and easy to set up, but Why: It improves basic can’t handle surprises automation by learning Why: They adapt to or new situations. and adapting to some new information and situations on their own. Examples: Best for complexity while still Examples: Perfect for repetitive tasks like following rules. researching sales leads sending emails after by evaluating multiple Examples: Useful for sign-ups or updating sources and prioritizing handling customer records with fixed contacts based on messages, sorting steps. relevance, or for requests by monitoring social media importance, or trends to identify recommending emerging topics and products based on deciding how to behavior. respond in real time.
Let’s have a look at what you can do with agentic automation in more detail.
Continuetothewrapupforthisunit
Wrap up
In Make, you can use three types of automation, each for a different purpose. Traditional automation handles simple, rule-based tasks that repeat the same way every time. AI automation takes this further by adding intelligence, using AI to analyze data, make decisions, or generate content. Agentic automation goes one step beyond, allowing AI agents to act independently, adapt to changing situations, and choose the best actions to achieve a goal.
Traditional automation depends on Boolean logic to follow fixed rules and get consistent results. AI automation combines Boolean and fuzzy logic, using both clear conditions and adaptable reasoning to handle complex or uncertain data. Agentic automation relies mainly on fuzzy logic, allowing it to adjust, learn, and make independent decisions in dynamic situations.
Each type of automation has its own strengths: traditional automation is fast and reliable for simple tasks, AI automation blends clear rules with flexibility to manage more complex situations, and agentic automation provides full autonomy to adapt to changing conditions.
Unit complete!
Well done!
You have completed the first unit! In the next unit, you’ll learn when it’s best to use agentic automation.
Mark this task complete to continue to the next unit.