I have been experimenting with AI tools since well before OpenAI launched ChatGPT. Back then, they were clunky and fragmented, little pieces of automation stitched together, producing more noise than clarity. To get real insights, I often had to lean on a different kind of intelligence: my wife. She was the unofficial co-pilot of Accolade Coaching, helping me think through problems, spot flaws in reasoning, and ask the uncomfortable questions that software simply couldn’t.
Even in those early days, a question nagged at me:
“Could AI eventually disrupt my business?”
At Accolade Coaching, much of what I do involves helping innovation leaders and entrepreneurs sharpen ideas, stress-test strategies, and uncover opportunities. On the surface, those are “thinking jobs”, the kind of work you’d expect a machine to get better and better at. And if a tool could do it faster and cheaper, wouldn’t that threaten what I offer?
Fast forward to today. AI language models have matured at an incredible pace. They can brainstorm dozens of ideas in seconds, draft reports with impressive polish, and analyse information at a scale no human could match. In some ways, they already outperform the “old” me who relied on manual effort and borrowed brainpower.
But here’s the surprising discovery: the more I use these tools, the more convinced I become that they can’t replace the most important parts of what I do. AI can handle tasks, but it can’t carry trust. It can suggest strategies, but it can’t sit with a founder in their moment of doubt. It can mimic empathy, but it doesn’t share in the risk or the pride of seeing an idea come alive. In other words, there are still entire categories of “jobs” AI will never steal.
You just have to know where to look.
Table of Contents
- The Three Jobs to Be Done — A Quick Primer
- AI’s Sweet Spot: The Functional Job
- Where AI Falls Short: The Social Job
- Where AI Falls Short: The Emotional Job
- Why the Last Two Jobs Matter Most
The Three Jobs to Be Done — A Quick Primer
To understand why AI falls short, we need to look at a framework I use often with my clients at Accolade Coaching: Jobs to Be Done (JTBD).
The idea, first popularised by Harvard professor Clayton Christensen and later refined by thinkers like Strategyzer, is that people don’t just buy a product or service, they hire it to get a job done. And crucially, not all jobs are created equal.
Jobs typically fall into three categories:
Functional Jobs are the practical, task-focused ones. They’re about achieving an outcome as efficiently as possible. A pharma researcher, for example, may need to screen thousands of compounds for a viable drug candidate. In coaching, a founder might want to map their business model on a canvas. These are jobs with clear inputs and measurable outputs, the ones that look easiest for technology to take over.
Social Jobs are about how we want to be perceived by others. A scientist presenting data at a conference doesn’t just want accuracy; they want to be seen as credible and trustworthy. A startup founder pitching investors isn’t only explaining a model; they want to be recognised as visionary and bold. These jobs are about signalling competence, authority, or belonging.
Emotional Jobs are the most personal. They’re about how we want to feel, or avoid feeling, while we do the work. The regulatory lead preparing a submission wants to feel confident that nothing is missing. The entrepreneur staring down a risky pivot wants to feel supported, not alone. Emotional jobs run beneath the surface, yet they often determine whether people adopt, stick with, or abandon a solution.
Businesses that focus only on functional jobs often fall into the trap of saying,
“Our product works, why isn’t it selling?”
But those that understand the full spectrum of jobs create solutions that resonate more deeply. They don’t just work; they make customers feel something and help them be seen in the right light.
And this is where AI comes into the picture. Because while it’s already proving powerful at functional jobs, the other two, the social and the emotional, remain stubbornly human.
AI’s Sweet Spot: The Functional Job
When it comes to functional jobs, AI is already proving itself to be a powerful ally. These jobs are the most straightforward: clear tasks with defined inputs and measurable outputs. They’re the kind of work that often feels repetitive or data-heavy for humans, which is exactly where machines excel.
In pharmaceuticals, functional jobs are everywhere. AI can scan through massive chemical libraries in hours, a task that would take a research team months. It can model drug–target interactions before a single experiment is run, or condense a stack of research papers into a neat, digestible summary. These are jobs that demand speed and precision, and AI delivers both in spades.
The same dynamic plays out in innovation coaching. Imagine a client comes in with a half-formed idea. Within minutes, AI can generate a draft business model canvas, outline customer personas, and even surface potential competitors. It can process pages of customer interview transcripts and pull out recurring themes faster than any human analyst. These outputs don’t just save time, they give coaches and entrepreneurs a solid base from which to explore ideas further.
It’s no surprise that this is AI’s comfort zone. Functional jobs can often be boiled down into patterns: if you give it enough data, it can find the relationships and replicate the process. And in this lane, AI will only get faster, cheaper, and more accurate over time.
But here’s the limitation: completing the functional job doesn’t necessarily get the real job done. In pharma, a neat dataset won’t convince regulators if no one trusts the messenger. In coaching, a clever business model won’t inspire a team if the founder doesn’t believe in it or can’t carry it in front of investors. That’s because functional jobs are only one part of the story, the most replaceable part. The real staying power lies in the social and emotional jobs, the ones AI can’t touch.
If functional jobs are about what gets done, social jobs are about how we’re seen while doing it. They shape perception, reputation, and credibility. And in fields where trust and authority are everything, the social job can be more decisive than the technical outcome itself.
Take pharmaceuticals. A scientist presenting clinical trial data isn’t just communicating results. They’re embodying credibility. Regulators, investors, and peers are scanning for signals: does this person inspire confidence? Do they speak with authority? Do they carry a track record of rigor and ethics that makes their conclusions trustworthy? AI can generate flawless graphs or tidy reports, but it cannot walk into that room with years of credibility attached to its name. Trust in pharma isn’t built in a dataset; it’s built through lived expertise, accountability, and reputation.
Innovation coaching has a parallel dynamic. When a coach steps into a boardroom to help a leadership team navigate a risky decision, the room isn’t only evaluating the ideas being discussed. They’re also evaluating the coach: do they understand our context? Can they handle our politics? Do they command enough respect to challenge the CEO when it matters? AI can flood the table with clever strategies, but it cannot earn that respect. It can’t sit across from a sceptical executive and, through presence and history, persuade them to act.
Social jobs often live in the subtleties — the pause before answering a difficult question, the body language that signals calm under pressure, the careful choice of words that makes an idea land without sparking defensiveness. These aren’t just “outputs,” they’re relational cues. They work precisely because they come from humans who have reputations, histories, and skin in the game.
This is why AI will always struggle here. It can imitate the language of authority, but it cannot carry the weight of being known, trusted, or accountable. The social job is about connection, not computation — and that’s territory machines simply can’t occupy.
Where AI Falls Short: The Emotional Job
If social jobs are about how others see us, emotional jobs are about how we feel while doing the work. They are the quiet but powerful drivers of human behaviour: the need for confidence, the relief of reassurance, the pride of achievement, and the fear of failure. Unlike functional jobs, these don’t show up in spreadsheets. Yet they often determine whether someone embraces a solution, sticks with it, or abandons it altogether.
In pharmaceuticals, emotional jobs surface constantly. A regulatory lead finalising a submission doesn’t just need the forms to be correct — they want to feel certain that nothing has slipped through the cracks. A clinical researcher chasing promising early data isn’t only collecting results — they want reassurance that years of effort may finally lead to a breakthrough. Even the act of sharing results carries an emotional weight: pride in contributing to science, relief that the long nights in the lab mattered. AI can crunch the data and highlight the patterns, but it cannot share in that pride, nor offer the calming presence of a trusted colleague who says, “We’ve got this.”
The same is true in innovation coaching. When a founder faces the possibility of pivoting their business model, the hard part isn’t generating options — it’s wrestling with the anxiety of letting go of their original vision. In these moments, a coach’s job isn’t simply to provide frameworks. It’s to sense when the client needs encouragement more than analysis, or when silence speaks louder than another strategy slide. Emotional jobs are fulfilled through empathy, trust, and human connection — not just information. AI might simulate encouragement, but it cannot mean it.
What makes emotional jobs unique is that they’re rooted in shared humanity. They require someone who feels the weight of the stakes, who carries responsibility alongside you, who celebrates or mourns with you when the outcome arrives. AI can mimic the words, but not the presence. And in moments of risk, loss, or triumph, presence is everything.
This is why emotional jobs — like social ones — remain beyond the reach of machines. Functional jobs may be automated, but the jobs of comfort, reassurance, and courage are human to their core.
Why the Last Two Jobs Matter Most
At Accolade Coaching, I work with many talented entrepreneurs and leaders who excel at functional jobs. They know how to build a working prototype, craft a clean financial model, or assemble a technically sound business case. I’ve seen founders who can spin up a product demo in record time, or scientists who can explain complex processes with crystal clarity. These skills are impressive, and they often form the backbone of a strong idea.
But here’s the catch: functional jobs alone aren’t enough. Too often, I meet innovators who have designed a solution that works beautifully on paper, but they’ve overlooked the other two jobs, the social and the emotional. They forget that customers don’t just “hire” a solution because it functions correctly. They hire it because it helps them look better in the eyes of others, and because it makes them feel safer, prouder, or more confident.
And this is where real value lies. If you can create a solution that doesn’t just perform as expected but also fulfills the social and emotional jobs of your customer, you’re no longer competing on features or price. You’re delivering something deeper, a product or service people want to identify with, advocate for, and yes, even pay a premium to own.
Think about the products you’ve personally been loyal to. Chances are, they didn’t just “work”, they made you feel something, and they changed how others saw you. That’s the territory AI can’t reach, and it’s where human ingenuity still shines brightest.
For businesses and entrepreneurs, this should be both a warning and an opportunity. A warning, because if you stay stuck in the functional lane, AI will soon overtake you. But an opportunity, because if you design with all three jobs in mind, you’re far more likely to create offerings that are not only well received, but that customers are willing to pay almost anything for.
Sources:
We didn’t do anything Wrong. . .[But somehow we lost]. (n.d.). Infobwana.
Statista. (2025, July 1). Global smartphone unit shipments 2009-2024.
Mallaband, B. (n.d.). The design tricks that made the Nokia 3310 world-beating. The Conversation.
Hill, A. (2013, February 26). Nokia: From ‘Burning platform’ to a slimmer management model. CNBC.
Minds, B. (2018, Jul 24). Why did Nokia fail and what can you learn from it? Medium.
