𝗕𝗼𝗮𝗿𝗱𝗿𝗼𝗼𝗺 𝗔𝗜: 𝗠𝗼𝗻𝗱𝗮𝘆 𝗥𝗲𝗮𝗹𝗶𝘁𝘆 𝗖𝗵𝗲𝗰𝗸 | 11
There’s a meeting that happens in most organizations sometime around month three of their AI journey.
The initiative has been announced. The budget has been approved. The strategy deck has been presented and applauded. And now someone asks the question that always comes next—the one that feels like the most natural, most responsible follow-up to everything that’s just been committed to.
“So who are we hiring?”
And the search begins. Job descriptions get written. Recruiters get briefed. LinkedIn gets searched. The organization starts looking for a Chief AI Officer, or a Head of Machine Learning, or a Senior Data Scientist, or some combination of all three—because surely, the thinking goes, the right people will know what to do with the strategy and the budget and the mandate.
Six months later, the right people have been hired. They’re smart. They’re experienced. Their CVs are genuinely impressive. And they’re sitting in an organization that still hasn’t answered the questions that needed answering before anyone was hired.
What problem are we actually trying to solve with AI?
Where in the business would AI create the most value?
What does success look like—specifically, measurably, in a way that connects to something the business cares about?
What data do we have, and is it in a state that makes any of this possible?
The talent didn’t fix the clarity problem. It inherited it.
And now the organization has an expensive team sitting on top of an unanswered question—producing activity, generating output, running pilots, and building prototypes—without the strategic foundation that would make any of it add up to something real.
This is the skills mirage. And it’s one of the most consistent and costly mistakes in enterprise AI right now.
Why hiring feels like the answer
I want to start here, because the instinct to hire is not irrational. It comes from a genuine and understandable place.
AI is new enough—and complex enough—that most leadership teams feel a real gap between what they know and what they need to know to make good decisions. The technology is moving fast. The landscape is confusing. The vendors are speaking a language that doesn’t always translate cleanly into business terms. And in that environment, the appeal of hiring someone who speaks that language fluently—someone who can navigate the complexity, make sense of the options, and tell leadership what to do—is completely understandable.
Hiring has also worked historically as a response to capability gaps. When organizations needed to get serious about digital marketing, they hired digital marketers. When they needed to build mobile experiences, they hired mobile developers. When they needed to think about cybersecurity, they hired security specialists. The pattern is deeply embedded—identify a gap, and hire to fill it.
The problem is that those capability gaps had relatively clear definitions. A digital marketer knows what digital marketing is. A mobile developer knows what mobile development requires. The role has a shape, and the person fills it.
AI doesn’t have that kind of clarity yet—not at the organizational strategy level, where the decisions about what to do with AI actually get made. And hiring into that ambiguity doesn’t resolve it. It just adds headcount to it.
What the job descriptions reveal
Here’s a practical test that I find genuinely revealing.
Pull up the AI-related job descriptions your organization has posted in the last twelve months—or the ones you’re about to post—and read them carefully. Not for the technical requirements. For what they reveal about the organization's understanding of what it actually needs.
What you’ll often find is one of three patterns, each of which tells a different story about the underlying clarity problem.
The impossibly broad role. This is the job description that asks for someone who can develop AI strategy, build and train models, work with business stakeholders, manage data infrastructure, oversee compliance, and present to the board. It’s asking for five different roles in one person—because the organization hasn’t yet figured out which of those five things it actually needs most urgently. The breadth of the requirement is a direct reflection of the narrowness of the strategic thinking.
The technically precise but strategically empty role. This one knows exactly what technical skills it wants—specific frameworks, specific model architectures, specific tools—but is completely vague about what those skills will be applied to. “Drive AI initiatives across the organization." “Identify and implement AI opportunities.” “Lead the company’s AI transformation.” These sound like job descriptions. They’re actually placeholders for the strategic conversation the organization hasn’t had yet.
The role is really a different role. Sometimes what the organization actually needs is a business analyst who can bridge between technical teams and business stakeholders. Or a data engineer who can fix the pipeline problems that are blocking everything else. Or a change management specialist who can help the organization actually adopt the tools that already exist. But because the conversation is framed around AI talent, the search is for AI talent—and the actual bottleneck never gets addressed.
Each of these patterns points to the same underlying issue. The organization is trying to hire its way to clarity rather than achieving clarity before it hires.
What happens when talented people land in unclear environments?
This is the part that I find most difficult to watch, because it does real damage—to the individuals involved, to the organization's AI credibility, and to the talent pipeline that every organization will need over the next decade.
Talented AI people—genuinely skilled, genuinely motivated, and genuinely capable of building things that matter—land in organizations that haven’t done the strategic work. And here’s what their first six months typically look like.
They spend the first month trying to understand the landscape. Meeting stakeholders. Reviewing existing systems. Trying to figure out what the organization actually has, what it actually needs, and where the real opportunities are. This is useful work. But it’s work the organization should have done before hiring them.
Get Yash Naik’s stories in your inbox
Join Medium for free to get updates from this writer.
They spend the next two months navigating competing priorities. Every department head has a different idea about where AI should go first. The sales team wants a forecasting tool. Marketing wants content automation. Operations wants process optimization. Finance wants anomaly detection. There’s no framework for evaluating these competing requests—because the strategic clarity that would provide that framework doesn’t exist. So the AI team either picks one arbitrarily, tries to do all of them superficially, or gets stuck in stakeholder management while producing very little.
By month four or five, something shifts. The talented person—who came in with genuine energy and genuine capability—starts to feel the friction of the environment they’ve landed in. They’re spending more time in alignment meetings than building anything. The work that does get done doesn’t connect to clear outcomes. There’s no way to know if what they’re doing is working because nobody defined what working looks like.
And then one of two things happens. They leave—taking their capability, their institutional knowledge, and their professional network with them—and the organization posts the same job description again, convinced the next hire will be better. Or they stay and adapt to the environment—producing the kind of safe, low-impact, easy-to-justify work that an unclear environment rewards, rather than the high-impact work that an unclear environment makes impossible.
Neither outcome is the talented person’s fault. The environment was the problem. The hiring just made the problem more expensive.
The clarity that needs to come first
Let me be specific about what clarity actually looks like before a hiring decision gets made—because I think it’s worth being concrete here rather than just gesturing at the concept.
A defined problem, not a defined technology. The starting point is not “we want to use AI for X.” It’s “we have a specific, costly, measurable business problem—and we believe AI is the most appropriate solution to it.” The difference matters enormously, because one leads to technology looking for a problem and the other leads to a solution looking for the right implementation.
An honest data assessment. Before any AI role gets posted, the organization needs an honest picture of its data landscape. What data exists? Where does it live? How clean is it? How accessible is it? Is there enough of it, in the right form, to support the AI applications being considered? These questions don’t require an AI specialist to answer—they require honesty and a bit of investigative work. But they completely determine what kind of AI talent the organization actually needs.
A success definition that means something. What would the organization be able to do in twelve months—specifically—that it can’t do today? What metric would move, by how much, and by when? Who is accountable for that outcome? If these questions don’t have clear answers before the hire is made, the person being hired has no foundation to build on.
An organizational readiness assessment. Is there executive sponsorship for this work—real sponsorship, not rhetorical support? Are the business stakeholders who need to engage with AI initiatives actually available and aligned? Is there a pathway from proof of concept to production—or will every successful experiment hit the same deployment bottleneck that previous ones did? The most talented AI person in the world cannot create organizational readiness that doesn’t exist. They can only work within it.
Why the best people leave—and what that does long-term
There’s a compounding dimension to the skills mirage that rarely gets discussed in the hiring conversation.
The AI talent market is competitive. The people with the skills organisations most want to hire have options — good options, with organisations that have done the strategic work and can offer an environment where their skills will actually be used effectively.
When those people join an unclear organisation and encounter the reality described above, they leave. Usually within twelve to eighteen months. And when they leave, they don’t just take their skills — they take their honest assessment of the organisation with them. They tell their network. The next hire is harder to make. The salary expectation goes up. The organisation develops a quiet reputation as a place where AI talent goes to be frustrated.
Meanwhile, the organisations that defined the problem before they hired the person — that gave their AI talent a clear mandate, real data to work with, genuine executive support, and a measurable outcome to build toward — are the ones retaining their best people. And those people are compounding. Building institutional knowledge. Growing internal capability. Creating the kind of AI advantage that’s genuinely hard to replicate.
The clarity gap doesn’t just affect the first hire. It shapes the organisation’s entire trajectory through the AI talent market for years.
The questions leadership needs to answer first
Before the next AI role gets posted, before the next recruiter gets briefed, before the next LinkedIn search begins — sit with these honestly:
→ What specific business problem are we hiring this person to solve — and can we describe it without using the word “AI”?
→ What does our data landscape actually look like — and have we been honest about what that means for what’s achievable?
→ What does success look like in twelve months — specifically, measurably, connected to something that matters?
→ Is the organisational environment ready to support someone doing this work — or are we hiring into a set of conditions that will frustrate them out the door?
If those questions don’t have clear answers — the hiring conversation is premature. The strategy conversation needs to happen first.
The Monday Reality Check
The AI talent shortage is real. The competition for skilled people is genuine. The organisations that build strong AI capability will have a meaningful advantage over the ones that don’t.
But the talent shortage is not what’s holding most organisations back right now. What’s holding them back is strategic ambiguity — the absence of clear answers to the questions that need to be answered before any hiring decision makes sense.
You cannot hire your way out of a clarity problem. The best data scientist in the world cannot tell you which business problem AI should solve first. The most experienced ML engineer cannot define success criteria that the organisation hasn’t defined for itself. The sharpest AI product manager cannot create executive alignment that doesn’t exist.
Clarity comes first. Hiring comes second.
Define the problem. Assess the data. Define success. Understand the environment. And then — with those foundations in place — hire the person who can build on them.
That person will stay. They’ll produce. They’ll compound.
And the skills mirage will quietly stop being a problem — because the organisation will finally have something real to build on.
💬 Have you watched genuinely talented people struggle in an unclear AI environment — or left one yourself? What was the turning point? I’d genuinely like to hear.
Boardroom AI: Monday Reality Check is a weekly series for leaders who want the honest version of the AI conversation — not the pitch deck version. If this resonated, share it with someone who needs to read it before the next AI role gets posted.
