The Talent Will Leave Before the Startups Arrive

If you're a large organisation with real distribution advantages, embedded client relationships, platform status, and brand recognition, you're probably feeling pretty good about AI.

And you should. Those moats still work.

AI doesn't automatically erode the advantages you've built over decades. Your bundled offerings still create switching costs. Your brand still signals trust. Your scale still lets you do things smaller players can't.

But here's what should keep you up at night: the threat isn't that startups will disrupt you. The threat is that you'll slowly lose the ability to do anything new.

The startups won't kill you directly. They'll nibble at the edges whilst your best people walk out the door. And by the time you notice the pattern, the capability to respond will have atrophied beyond recovery.

This is how giants die. Not with a bang, but with a slow fade into irrelevance.

The Ankle Biters Are Getting Faster

Let's be honest about what AI has changed for your competitors.

A startup today can do things that would have required a 50-person team five years ago. Three people with the right AI tools can produce marketing campaigns, software prototypes, customer research, and go-to-market strategies at a pace and quality that used to require significant headcount.

They're not better than you. They're faster. And they're cheap enough to take risks you can't justify.

Each individual startup isn't a threat. But there are a lot of them. And they're all looking at the edges of your business—the underserved segments, the features you never prioritised, the customer needs you decided weren't worth the effort.

Death by a thousand cuts doesn't require any single cut to be fatal.

What AI has done is accelerate the metabolism of these small players. They iterate faster. They ship faster. They learn faster. The cycle time from idea to market has compressed dramatically.

Meanwhile, your planning cycle is still annual. Your approval process still requires seven signatures. Your technology decisions still go through procurement. By the time you've decided to respond to a competitive threat, three more have emerged.

This isn't a criticism. It's physics. Large organisations have coordination costs that small ones don't. That's always been true.

What's changed is the speed differential. The gap between how fast you can move and how fast they can move has widened. And it's widening further every month as AI tools improve.

Your Best People Can See the Future

Here's the part that should really worry you.

Your most capable people—the ones with the best judgment, the most developed taste, the deepest expertise—they can see what's happening. They read the same articles you do. They understand AI at least as well as you do, probably better.

And they're doing the maths.

They're looking at your organisation and asking: Can this place adapt fast enough? Will I be able to do meaningful work here? Or will I spend the next five years fighting bureaucracy whilst the interesting stuff happens elsewhere?

For many of your best people, the answer is increasingly clear. They can move faster somewhere else. They can have more impact somewhere else. They can learn more somewhere else.

Startups are attractive to talent not because they pay more (they often don't) but because they offer speed and agency. The ability to try things. The ability to ship. The ability to see the results of your decisions without waiting for quarterly reviews.

When your best people leave, they take something with them that's very hard to replace: judgment.

Remember the three-layer model. AI has made cognitive production cheap. What's scarce now is the judgment to evaluate that production, the accountability to own decisions, and the taste to distinguish excellent from merely acceptable.

Those capabilities are embedded in your experienced people. The ones who "just know" what good looks like. The ones who can make the call when the data doesn't give a clear answer. The ones who have earnt the right to be accountable through years of demonstrated judgment.

When they leave, you don't just lose headcount. You lose the capability that actually matters.

The Kmart Lesson

Let me tell you a story about a company that had every advantage and still lost. This is one I learnt about back when studying at University and never forgot.

Kmart in the 1980s was a retail giant. They had stores everywhere. They had brand recognition. They had supply chain infrastructure. They had everything that should have protected them from competition. Somehow they are still a thing here in Australia.

And then Walmart came along.

Walmart didn't beat Kmart by being dramatically different. They beat them by being relentlessly better at the fundamentals. Better logistics. Better inventory management. Better cost structure. Better execution, day after day, store after store.

Kmart could see what Walmart was doing. It wasn't a secret. But they couldn't respond effectively because they had lost the internal capability to execute at that level. Too many years of coasting on their market position. Too many years of promoting the wrong people. Too many years of underinvesting in the operational excellence that actually mattered.

By the time the threat was obvious, the muscle had atrophied. They couldn't just decide to be operationally excellent. That capability had to be built over years, and they hadn't built it.

This is the pattern that should terrify large organisations today.

Your moat protects you from direct disruption. Startups can't just walk in and take your customers. But your moat doesn't protect you from losing the ability to innovate internally. It doesn't protect you from the slow erosion of capability that happens when your best people leave and aren't replaced. It doesn't protect you from becoming Kmart.

The Capability Crisis Inside Large Organisations

Here's what I see when I work with large enterprises.

The senior leaders who understood how things actually work are retiring. They're not being replaced with people of equivalent capability because those people are expensive and the organisation optimised for cost.

The middle managers who could translate between strategy and execution have been squeezed out. Layers were cut, but the work of translation didn't disappear. It just stopped happening effectively.

The technical experts who could evaluate vendor claims and guide technology decisions have left for startups or consulting. The organisation now depends on vendors to tell them what they need, which creates obvious problems.

The institutional knowledge about why things are done a certain way has been lost. Decisions that made sense in context now continue as policy even though the context has changed.

What remains is process. Lots of process. Approval workflows and governance frameworks and risk management procedures. All of which made sense at some point, but which now primarily serve to slow everything down.

This is the capability crisis. Not a lack of tools or technology or budget. A lack of the human judgment required to do anything new.

And AI makes it worse before it makes it better.

Because AI produces abundant cognitive output. More options to evaluate. More possibilities to consider. More decisions to make. If you don't have the judgment capacity to handle that abundance, you just drown in it.

Buying more AI tools when you lack the capability to use them effectively is like buying a faster car when you've forgotten how to drive.

Why Operational Change Has to Come First

The typical large-organisation response to competitive pressure is to launch initiatives. Innovation labs. Digital transformation programmes. AI centres of excellence. Partnerships with startups. Hackathons.

These are not bad things. But they're not operational change. They're innovation theatre.

Real operational change means changing how the core business actually works. How decisions get made. How resources get allocated. How people get promoted. How success gets measured.

That's much harder than launching an innovation lab. And it's the only thing that actually matters.

Let me be specific about what operational change looks like in the context of AI.

It means giving people permission to use AI tools in their actual work. Not in sandbox environments or approved use cases, but in the real day-to-day work that matters. This requires accepting some risk. It requires trusting your people. It requires moving faster than your legal and compliance teams are comfortable with.

It means changing how you measure productivity. If you're still measuring output volume, you're measuring the wrong thing. AI makes output volume trivially achievable. What matters now is the quality of judgment. Are people making better decisions? Are they catching things that would have been missed? Are they adding value that AI cannot add?

It means restructuring teams around judgment rather than production. You need fewer people doing the producing and more people doing the evaluating. That's a painful transition because it means admitting that some roles have become less valuable. But pretending otherwise doesn't make it less true.

It means investing in developing judgment, accountability, and taste. These capabilities don't appear automatically. They have to be cultivated through experience, mentorship, and deliberate practice. If you're not actively developing these capabilities in your people, you're falling behind.

It means making decisions faster. The coordination costs of large organisations are real, but many of them are self-imposed. Every approval chain that doesn't add value, every committee that exists because it's always existed, every process that slows things down without reducing real risk—these are choices. You can choose differently.

Retaining Talent in an AI World

Your best people want to work somewhere that's going somewhere.

That sounds obvious, but think about what it actually means. They want to see progress. They want to ship things. They want to learn and grow. They want their judgment to matter.

If your organisation feels stuck, if nothing changes, if the same problems persist year after year despite endless discussion, your best people will leave. Not because they're disloyal, but because they're rational.

Here's what retaining talent actually requires:

Show them you're serious about change. Not with announcements and initiatives, but with actual changes to how things work. Kill a process. Eliminate an approval layer. Ship something in weeks instead of months. Demonstrate that change is possible.

Give them agency. The most capable people want to make decisions, not follow procedures. If every decision requires escalation, if every risk requires a committee, if every idea requires a business case, you're driving away exactly the people you need most.

Invest in their development. Not with training programmes that check boxes, but with real capability building. Pair them with mentors who have great judgment. Give them problems that stretch them. Create feedback loops that help them improve.

Let them use modern tools. Nothing signals organisational stagnation like forcing people to use outdated technology. If your people can't use AI tools that their friends at startups use freely, they'll wonder why they're working for you.

Connect their work to meaning. Your organisation presumably exists to do something worthwhile in the world. Make sure your people can see that connection. Make sure they know their judgment and effort matter.

None of this is revolutionary. It's just treating capable people like capable people. But it requires operational change, not just rhetorical change.

The 90-Day Test

Here's a challenge.

Pick one thing your organisation does that currently takes six months or more. A product launch, a technology implementation, a market entry, whatever.

Figure out how to do it in 90 days.

Not by cutting corners or accepting lower quality. By actually changing how you work. By removing unnecessary approvals. By empowering people to make decisions. By using AI to accelerate the parts that can be accelerated. By accepting that speed requires trusting your people.

If you can do this, if you can demonstrate that your organisation can move at a different pace, you'll learn something important about what's possible. And your best people will notice.

If you can't do it, if every attempt runs into immovable obstacles of process and politics and "that's not how we do things here," you'll learn something important about your organisation's capacity for change.

Better to learn that now than after another 20% of your best people have left.

The Path Forward

Let me be direct about what's at stake.

Your distribution moat is real. Your brand is real. Your scale is real. These advantages aren't going away tomorrow.

But advantages erode when you stop earning them. When you stop innovating. When you stop attracting and retaining the best people. When you lose the capability to respond to a changing market.

AI accelerates everything. It accelerates the startups nipping at your edges. It accelerates your best people's realisation that they could move faster elsewhere. It accelerates the gap between organisations that can change and organisations that can't.

The organisations that thrive will be those that use this moment to rebuild operational capability. Not to launch initiatives, but to actually change how they work. To push decision-making down. To develop judgment at every level. To create environments where capable people want to stay and grow.

The organisations that decline will be those that mistake their current advantages for permanent protection. That respond to competitive pressure with theatre instead of change. That lose their best people one by one whilst reassuring themselves that the moat still holds.

The startups aren't coming for you directly. But your talent is leaving, your capability is eroding, and the edges of your empire are being nibbled away.

The question isn't whether you can withstand disruption. It's whether you can change fast enough to remain worth working for.

The talent will leave before the startups arrive. By then, it's too late.

This is the fourth in a series exploring how AI is reshaping capability. Previously: "Judgment, Accountability, and Taste", "From Production to Judgment", and "The Danger of 'We Produce Cognitive Work.'"

If you're a leader who recognises these patterns in your organisation and wants to build real AI capability, not just buy tools, the AI Capability Intensive Executive Track is designed for exactly this challenge. Four weeks of building capability that compounds.