Retail and ecommerce teams using AI for customer service are resolving 40-60% more tickets without more staff, cutting cost-per-ticket by 30%+, and handling seasonal spikes 3x faster.

But here's what separates winners from everyone else: they started with the data, not the hype.

Gladly handles the predictable volume, FAQs, routing, returns, order status, while your team focuses on customers who need a human touch. The result? Better experiences. Lower costs. Real competitive advantage. Ready to see what's possible for your business?

See it in action

Listen, we need to talk about the AI job apocalypse everyone keeps predicting. Because Microsoft just dropped a study that actually looked at what AI is doing in the real world, and the results are way more interesting than the "robots are coming for your job" narrative you've been fed.

Here's what they did: Microsoft analyzed 200,000 actual conversations people had with Bing Copilot over nine months in 2024. Not speculation, not theoretical models—actual usage data. They mapped these interactions to real job tasks from the Labor Department's O*NET database and created an "AI applicability score" for different occupations. Higher score means AI can handle more of that job's tasks. Lower score means AI is basically useless.

The TLDR: If your job involves a keyboard and manipulating information, AI is coming for parts of it. If your job involves getting your hands dirty, breaking a sweat, or looking another human in the eye while doing something physical? You're golden.

The knowledge workers are, frankly, cooked—or at least facing the most disruption. Translators, writers, journalists, historians, salespeople, customer service reps. These roles scored highest on AI applicability because they involve exactly what current AI excels at: gathering information, drafting text, summarizing, answering inquiries.

But here's the kicker: the jobs at the bottom of the list—the 40 occupations least affected by AI—tell us everything about what makes work fundamentally human.

Top of the "AI can't touch this" list? Dredge operators. People who operate dredging equipment to excavate waterways scored a literal zero on AI applicability. Same with bridge and lock tenders and water treatment plant operators. Zero. AI couldn't help with a single core task.

Why? Because these jobs exist in the messy, unpredictable physical world where you need hands, spatial awareness, real-time judgment, and the ability to respond to conditions that change minute by minute.

The pattern holds across the entire list of 40 jobs. Let me give you the full picture:

Healthcare support roles dominate: phlebotomists drawing blood, nursing assistants providing patient care, surgical assistants in the OR, ophthalmic technicians conducting eye tests. You can't automate empathy, and you definitely can't automate the delicate motor control required to insert an IV or assist in surgery.

Skilled trades and maintenance: cement masons, roofers, tire builders, automotive glass installers, floor sanders, pile driver operators. These require tactile skills, problem-solving in three-dimensional space, and adapting to materials that don't behave consistently.

On-site service work: dishwashers, maids, massage therapists. Jobs where the whole point is physical presence and human touch.

Equipment operators: logging equipment operators, industrial truck drivers, paving machine operators, motorboat operators. Operating heavy machinery in dynamic environments where one wrong move could kill someone.

Even some highly trained professionals made the list—oral surgeons, prosthodontists. Because performing intricate procedures inside someone's mouth requires dexterity and judgment that AI simply cannot replicate.

The thing is, this isn't really about AI's limitations. It's about what makes work human.

Notice what all these protected jobs have in common? They require either:

AI is phenomenal at pattern recognition in digital spaces. It can process information, generate text, analyze data. But it's completely lost when you need to decide whether that concrete is setting properly, comfort a scared patient, or adjust your technique because the wood grain isn't cooperating.

Microsoft's lead researcher Kiran Tomlinson put it perfectly: the goal is to show how AI "might change how work is done—not necessarily replace jobs." Even in the most AI-exposed occupations, they found no job where AI could handle 100% of the tasks.

We've been having the wrong conversation about AI and jobs. The debate shouldn't be "will AI take my job?" but rather "which parts of my job can AI handle, and how do I evolve the rest?"

For knowledge workers, this means accepting that the routine information-processing parts of your role—the research, the first drafts, the data compilation—are increasingly AI territory. But the creative synthesis, the strategic thinking, the relationship management? Still yours.

For the people doing physical, hands-on work? This is actually validation. For decades, we've been told that white-collar knowledge work is the future and manual labor is being left behind. Turns out, in the AI age, the equation might flip. The jobs we've undervalued—the ones requiring human presence, physical skill, and real-world judgment—are the most AI-resistant.

A first-line supervisor of firefighters made the list. Think about that. In an emergency, when lives are on the line and conditions are chaotic, you need a human making split-second decisions. No amount of machine learning can replicate that.

This study accidentally reveals something profound about human capability: our bodies and our embodied cognition are way more sophisticated than we give them credit for.

We've spent the past 20 years convinced that information work is the pinnacle of human achievement. But a phlebotomist finding a vein, a roofer assessing structural integrity, a massage therapist reading muscle tension—these require integration of sensory input, motor control, and judgment that we're nowhere close to replicating artificially.

The jobs AI can't touch aren't "lesser" jobs. They're jobs that require the full human package: physical intelligence, emotional awareness, real-time adaptation to an analog world that doesn't follow clean digital rules.

The future isn't robots replacing humans. It's humans working alongside AI in domains where that makes sense, while an entire ecosystem of physical, hands-on, people-centric work remains fundamentally human.

If you're a writer, yes, you'll need to figure out how to use AI tools. But if you're a cement mason? Keep perfecting that trowel technique. If you're a nursing assistant? Your job security is better than most tech workers right now.

The real question isn't whether AI will take jobs. It's whether we'll finally start valuing the deeply human work—the physical, the embodied, the present—that keeps society actually functioning while everyone else is busy typing at computers.

Because at the end of the day, AI can write this newsletter. But it can't fix your roof, draw your blood, or make sure the bridge opens when the ship needs to pass through.

And that's not a limitation. That's the whole point.

According to Microsoft’s research, the following 40 occupations scored lowest on AI applicability – in other words, these are the jobs least threatened by generative AI as of now. They rely heavily on manual skills, in-person interaction, or complex real-world conditions that AI tools struggle to replicate. (By contrast, jobs ranking higher on AI overlap – like those involving writing and data – are not in this list.) Below are the 40 AI-“safe” roles identified, ranked from lower impact to virtually no AI impact: