Teaching AI to Cook—But It’s an Alien Chef

Optimise that omelette.

In the world of AI-driven system design, we often hear that AI operates like an “alien mind”—approaching problems in ways that are starkly different from human thinking. This metaphor isn’t just provocative; it’s a window into understanding how AI could fundamentally change the way we approach the design of automated workflows for small and medium-sized enterprises (SMEs) in B2B markets.

But let’s put it this way: imagine you’re trying to teach a friend how to cook. Except this friend is not your average sous-chef—they’re from another planet. They’re an alien chef who can process ingredients faster than you, measure with perfect precision, and follow a recipe to the exact gram. Sounds ideal, right? Well, not exactly. Their taste buds don’t work like yours, and their idea of “delicious” might involve combinations you’d never dream of—like mixing garlic with chocolate. Sure, they can chop onions faster than you, but would you trust them to come up with a perfect three-course meal?

This is exactly how we work with AI in system design. The AI has the computational power to handle vast amounts of data and follow a strict set of rules, but its solutions can often feel… a bit strange. Like the alien chef, it may suggest workflow “recipes” that are efficient but don’t always make sense to human users. It can uncover creative combinations and optimizations that we wouldn’t normally think of, but those solutions might need refining before they can work in a real-world business environment.

In this post, we’ll walk through how this alien approach to system design shapes our thinking in building AI-driven workflows for SMEs. We’ll explore how we take the AI’s raw ideas—the garlic-chocolate combos—and work to fine-tune them into practical, powerful solutions that balance efficiency with usability. Along the way, we’ll show how human designers collaborate with AI, creating a product that not only helps businesses run smoother but also makes sense for the people using it.

In traditional human-driven system design, we approach problems linearly, using our experiences and biases to inform solutions. AI, on the other hand, traverses vast conceptual spaces—paths that we, as humans, might never think to explore. Imagine our alien chef again. While you might methodically follow your recipe step by step—peeling, chopping, sautéing—the alien chef is busy using all the ingredients in the kitchen in ways you wouldn’t expect. They’re not bound by human logic; they’re experimenting, combining, and optimising in ways that make sense from their perspective but might seem baffling at first glance. And it’s these unconventional paths that can revolutionise how we think about workflow automation.

When designing our AI-driven software, we’re not just teaching the system to automate repetitive tasks like peeling carrots. We’re letting the AI chef uncover patterns and optimisations in how those carrots are chopped, cooked, and served that even the most seasoned human chefs (or operations managers) might miss.

Let’s take scheduling as an example. Human designers tend to organize data based on human structures—like hours, minutes, and seconds—because that’s how we’ve been trained to think. But AI doesn’t rely on these constructs. It might uncover that instead of scheduling work in 60-minute increments, certain tasks could be completed faster by optimizing based on workload patterns rather than rigid time slots. It might realise that “busyness” isn’t about how long a task takes but rather when people are available or unavailable, and optimise schedules in ways that account for gaps we don’t normally see.

For instance, where a human designer might think, “Let’s block off 30 minutes for this meeting,” the AI might look at patterns in workflow and say, “Actually, this task only ever takes 22 minutes when done in the afternoon. Let’s optimize around that.”

Similarly, in workflow automation, while we might create rigid steps for repetitive processes based on how we traditionally expect things to flow, AI will synthesise huge amounts of data and find efficiencies that humans wouldn’t naturally consider. The AI sees opportunities to cut out unnecessary steps or reorder tasks based on how the data is actually being used, much like how our alien chef might discover that adding the garlic after cooking the chocolate (despite sounding odd to us) actually makes for a smoother blend.

This ability to work beyond human-defined structures is where the “alien mind” metaphor becomes most valuable. AI uncovers solutions and efficiencies that aren’t bound by our biases—redefining not just how things should be done, but how they could be done in ways that are smarter, faster, and less obvious to the human eye.

Instead of viewing AI’s alien intelligence as a challenge, we see it as a tremendous advantage—much like the success of an innovative chain of food outlets that is constantly staying ahead of its competitors. Imagine this: a small but rapidly growing chain of restaurants starts using an alien chef to design their menus. While traditional competitors are sticking to the same tried-and-true recipes, this chain is allowing the alien chef to experiment with new ingredients, flavours, and cooking techniques. The AI chef’s combinations are sometimes a bit unconventional, but over time, it uncovers new dishes that are not only unique but also highly scalable across multiple locations.

What sets this chain apart is its ability to be adaptable. The alien chef quickly notices patterns in customer preferences at different locations. At their outlet in Glasgow, spicy dishes are popular, but in Edinburgh (obs), people prefer milder options. The AI synthesises all of this data and tailors the menu to match regional tastes, ensuring each outlet remains competitive and relevant to its customer base.

Similarly, by allowing our AI to experiment with workflow automation, we’re aiming to create solutions that help our SME clients optimise operations in ways that their competitors simply aren’t. Just like the restaurant chain that scales its alien chef’s insights across its outlets, our AI-driven software will identify patterns across different industries and business models and fine-tune workflows accordingly. It might be noticed that certain tasks are more efficient at specific times of the day or that certain workflows work better for businesses of a particular size or structure.

This adaptability and scalability are game-changers for businesses looking to streamline their operations. Our customers won’t just get a one-size-fits-all solution—they’ll get a dynamic workflow engine that evolves with their business, much like the restaurant chain that can tweak its menu based on ever-changing customer preferences. This ability to scale and refine workflows as their businesses grow is what will set our clients apart in their industries, allowing them to stay competitive, just like that successful chain of outlets keeping their customers coming back for more.

AI as an Explorer of Workflow Spaces:

Premise: AI can explore vast and unconventional spaces in workflow design, far beyond what human designers typically consider.

• Action: We allow AI to analyse vast amounts of real-world data from different SMEs, freeing it to identify inefficiencies, patterns, and optimisations that go beyond rigid human-defined workflows.

• Outcome: AI uncovers workflow solutions that are highly efficient and adaptable, even discovering processes that might seem counterintuitive at first but work in practice. For example, much like a chef discovering that certain ingredients are better prepared in a non-traditional order, the AI finds new ways to streamline operations that humans wouldn’t have thought to explore.

  
  

Human-AI Collaboration:

Premise: AI’s computational power allows it to rapidly create optimised workflows, but human designers add the needed context and practical insights.

• Action: Our team works alongside AI, using its data-driven insights to refine and adjust workflow designs, ensuring they’re not only efficient but also intuitive and suitable for the SMEs we serve.

• Outcome: The collaboration results in workflows that blend AI’s efficiency-focused solutions with human-friendly design principles, much like a chef refining a recipe based on customer feedback to make it more appealing while maintaining its originality. The result is a software product that scales efficiency without losing the human touch.

  
  

By embracing AI’s alien intelligence in our system design, we unlock the potential to create workflows that are smarter and more efficient than any purely human-driven design could achieve. However, we will always need to balance AI’s unfamiliar, data-driven outputs with human oversight to ensure usability, practicality, and industry relevance.

As we move through the next stages of our design process, we’ll be focusing on how AI generates these unconventional designs and testing them against human-led standards to ensure our workflow automation software is as groundbreaking as we envision.

This first post introduces the concept of AI’s alien mind and how its non-human approach will influence the way we build automated workflows for SMEs. In the upcoming posts, we will explore how we validate these AI-generated solutions, how to prioritize them, and eventually move toward building our MVP. Stay tuned!

Generative AI Space and the Mental Imagery of Alien Minds

How Metaphors Shape our Understanding of AI

Challenging The Myths of Generative AI

Artificial Intelligence and Metaphor