STORY, MEET STRATEGY

Let’s make marketing feel less robotic and more real.

Find resources that bring your message—and your business—to life.

Humans Beat AI at Adapting—Here’s Why That Matters for Your Business

Humans Beat AI at Adapting—Here’s Why That Matters for Your Business

September 03, 20258 min read
Custom HTML/CSS/JAVASCRIPT

By Vicky Sidler | Published 3 September 2025 at 12:00 GMT+2

If you’ve ever walked into a hotel bathroom in the dark and instinctively located the light switch, congratulations. You’ve already beaten artificial intelligence at something it can’t yet do.

According to a Harvard Business School study led by Julian De Freitas, humans outperform AI in a crucial area: adapting to new and unexpected situations. While machines may crunch data faster than we can blink, they still fall apart when the environment shifts—even just a little.

Let’s unpack what this means for small business owners trying to decide whether to lean into automation or keep a human in the loop.


TL;DR

  • Humans beat AI in all four flexibility tests designed to simulate real-world change

  • AI lacks “self-orientation”—the ability to understand its own role in a shifting situation

  • This gap means AI struggles with context and can’t pivot the way people do

  • For business owners, that’s a reminder to deploy AI wisely and keep humans involved

  • Managers should be especially careful with AI in fast-changing environments

Need help getting your message right? Download the 5-Minute Marketing Fix


The Secret Sauce—Human Self-Orientation:

Before you even start a task, your brain does something most of us never think about. It takes in your surroundings, figures out where you are, and reorients your actions based on the context.

This mental process—called “self-orienting”—helps you switch from reading an email to answering the front door, or from comforting a child to calculating payroll.

In a business setting, this is the difference between a staff member who notices a problem and fixes it, and a system that just sits there waiting for input.

AI, for all its flash, still doesn’t have this built in. According to De Freitas, “What makes humans so effective is that we can do many things. We’re pretty flexible.”

The Mario Kart Test—Humans 4, AI 0:

To test this, researchers designed four video game-like tasks. Imagine a stripped-down version of Mario Kart, where players control a red square and have to reach a goal. The catch? There are four red squares on the screen—but you can only control one.

Humans quickly figured out which square was theirs. Then they got to the goal. AI? Not so much. It had no real concept of “self,” so it couldn’t determine which square it controlled. The result: confusion, bad guesses, and poor performance.

Even advanced AI trained on frame-by-frame data struggled. It simply couldn’t reorient fast enough when conditions changed. As De Freitas put it, “Self-orientation doesn’t seem to exist at all for AI.”

This might sound like a problem for gamers, but it’s a major issue in business too.

Why It Matters for Small Business Owners:

Let’s say you run a home services company. A client calls and says their elderly father is locked in the bathroom. Your technician gets there and sees that the door handle is old and fragile.

In that moment, your technician doesn’t just follow a checklist. They adjust. They move slowly, speak gently, and take extra care not to worsen the situation.

AI tools—whether they’re bots, scheduling systems, or so-called smart assistants—can’t do this yet. They’re great at set routines. But when the task shifts from what they’ve seen before, they either freeze or fail.

For business owners, this means:

  • AI is best used in predictable tasks

  • Human judgement is still critical when things shift

  • Automation should always be monitored, especially in customer-facing roles

Klarna Tried to Replace Humans—And Paid the Price:

If you want a real-world example of AI failing to self-orient, look no further than Klarna. The fintech giant replaced 700 support agents with a chatbot, claiming it could handle millions of customer queries across 35 languages.

But when those customers needed empathy, nuance, or even basic dispute resolution, the bot failed hard. Satisfaction scores dropped 22 percent. Klarna had to backtrack and rehire humans.

The takeaway is simple: humans know when the problem has changed. Bots don’t.

Klarna wasn’t alone. Companies from fintech to fast food jumped on the “digital employee” train, hoping to cut costs. What they got instead were PR headaches and rehires.

These failures have one thing in common: overestimating what AI can do without human oversight.

Empathy, judgment, and creative problem-solving are still human strengths. Even the smartest bot can’t wing it when the rules change.

What AI Still Can’t Do:

According to De Freitas, even high-end AI needs to see huge amounts of data to “learn” how to respond. And even then, it’s guessing based on repetition, not real understanding.

That might work for spelling corrections or traffic updates. It does not work for high-stakes, fast-changing jobs like customer support, healthcare, or crisis management.

To quote De Freitas again, “The current way to achieve this feat with AI is to throw a lot of data at it and hope that AI sees everything it needs to see. But I don’t think that’s a flexible, fail-safe approach.”

When to Trust Automation—And When Not To:

Here’s the practical bit. If you’re considering new software, AI tools, or automation in your business, ask yourself these two questions:

1. Will the environment stay consistent?

AI tools work best when the rules don’t change. Think payroll, order processing, or basic scheduling.

2. Will this task need context or judgement?

If the answer is yes—such as adapting your tone to different clients, spotting subtle problems, or switching tasks midstream—then AI may struggle.

De Freitas puts it simply: “If you more deeply understand why your AI systems are limited, you are probably better equipped to know when and how to deploy them.”

What This Means for Marketing:

Marketing often lives in the grey zone between consistency and chaos. You want systems in place, but you also need to respond to shifting moods, platforms, and customer expectations.

Here’s what I recommend to my clients:

  • Use AI to support, not replace, your creative process: Let it help with first drafts, trend research, or formatting—but keep a human brain in the loop.

  • Build systems that are flexible: Email sequences, content calendars, and funnels should be adaptable. Don’t lock yourself into rigid automation you can’t quickly change.

  • Prioritize clarity in your message: If your core message is clear and strong, you’ll navigate changing conditions much more easily. You won’t be rewriting your homepage every time a trend shifts.

And if you’re not sure your message is clear, start with the 5-Minute Marketing Fix. It’ll walk you through crafting one sharp sentence that works everywhere—from social posts to sales pitches.

👉 Download it here

Keep Things In Perspective:

AI is not magic. It’s not a person. And it’s definitely not ready to deal with your crankiest customer on a rainy Monday.

It can speed things up, but it needs guardrails. It can assist, but it still can’t adapt like a human.

The smartest small businesses will treat AI like a clever intern—helpful, fast, but not ready to run the place solo.


Related Articles

Klarna Reverses AI Customer Support After Backlash—Why Human Help Is Back in 2025

This article provides the full story behind the Klarna example. It shows how a billion-dollar company had to rehire humans after its AI-only support tanked customer satisfaction.

AI Replacing Humans Backfires—What CEOs Miss

From botched experiments to legal risks, this piece explores what happens when businesses try to automate too much, too fast—and how to do it right instead.

AI Can't Replace Expertise—Tea Data Breach Proves It

If you’re worried about using AI in sensitive areas, this one’s for you. It unpacks how a dating app’s reliance on AI led to a massive security breach—and why skilled human review still matters.

AI Is Coming for White-Collar Jobs—and That's Just the Start

Want to understand where all this is headed? This article explores the bigger picture of AI in the workplace, and why the most successful professionals are learning to use AI instead of fearing it.

Vogue's AI Model Ad Backfires—Marketing Lessons for Small Businesses

This brand misstep shows what happens when AI replaces humans in creative campaigns. It’s a cautionary tale about authenticity, audience values, and knowing when to keep real people in the frame.

AI Copyright Lawsuit: What Small Businesses Must Know Now

If you’re using AI-generated content, read this next. Disney’s lawsuit against an AI firm could reshape what’s legal in marketing, content creation, and brand storytelling.


FAQs on Human Adaptability vs AI Limitations

What is “self-orienting” and why does it matter?

Self-orienting is the ability to understand your surroundings, recognise your role, and adjust your actions accordingly. It’s what allows humans to quickly adapt to new environments or tasks—something AI still struggles to do.

How did researchers test AI’s adaptability?

Harvard researchers designed video games where both humans and AI had to identify their digital avatar and reach a goal. While humans adapted quickly, AI consistently failed to figure out what it was controlling or how to respond when the game environment changed.

What real-world examples show AI’s limits?

Klarna’s failed rollout of AI-only customer support is a prime example. Customer satisfaction dropped by 22% when bots replaced human agents. The company later reversed course and brought human support back.

Can AI ever learn to self-orient like humans?

Maybe one day, but not yet. Current AI relies heavily on training data and struggles when faced with unfamiliar or changing situations. Researchers are trying to build more adaptable systems, but it’s a long road.

What tasks is AI actually good at?

AI excels at repetitive, structured, low-context tasks—like sorting emails, generating drafts, or answering basic customer questions. It should support humans, not replace them, in situations requiring nuance or flexibility.

How should small businesses approach AI tools?

Start small, test in safe areas, and keep people involved. Focus on automation for simple tasks, but make sure real humans are available when stakes are high or problems change unexpectedly.

How does this apply to small business marketing?

Marketing environments shift fast—so while AI can help with content drafts or scheduling, your core message needs to be human and clear. For help with that, download the5-Minute Marketing Fix. It’s free and will help you write one sentence that works across platforms.

blog author image

Vicky Sidler

Vicky Sidler is a seasoned journalist and StoryBrand Certified Guide with a knack for turning marketing confusion into crystal-clear messaging that actually works. Armed with years of experience and an almost suspiciously large collection of pens, she creates stories that connect on a human level.

Back to Blog
Strategic Marketing Tribe Logo

Is your Marketing Message so confusing even your own mom doesn’t get it? Let's clarify your message—so everyone wants to work with you!

StoryBrand Certified Guide Logo
StoryBrand Certified Guide Logo
Duct Tape Marketing Consultant Logo
Woman Owned Business Logo

Created with clarity (and coffee)

© 2025 Strategic Marketing Tribe. All rights reserved.

Privacy Policy | Terms of Service | Sitemap