Skip to content
Vinthony

AI career disruption

The job market is being restructured, slowly in some places and quickly in others. The useful question is rarely “will AI replace my job” — it's “which version of my role do I want to be, and what would I do this quarter to become it?”

The shape of disruption

The first wave of AI career disruption doesn't look like mass redundancies. It looks like task substitution. Particular activities that used to take a person two hours (drafting an email; summarising a 30-page document; producing a first-draft analysis; generating boilerplate code; doing basic customer triage) now take ten minutes with AI. Productivity per worker rises. For a while, that's good for the workers who adopt the tools fast.

The second wave is structural. Companies notice that the same headcount is producing more, or that fewer people produce the same. Hiring slows. Specific roles get re-scoped. Some get eliminated. This wave is harder to spot because it's happening at the strategy level, not the task level.

We're currently in the late first wave and early second wave for most knowledge work. The companies that have already restructured (and many have, quietly) tell us what the next 24 months look like for the rest.

Three versions of every role

In any role being disrupted, three versions emerge over 2-3 years.

  1. AI-augmented. The person who uses AI tools well, ships more in less time, and re-invests the saved hours in higher-judgement work. Often promoted; often expands scope.
  2. AI-substituted. The version of the role that was mostly the routine, automatable part. Hours per task drop until the headcount math no longer works. Restructure follows.
  3. AI-defensible. The version concentrated on judgement, relationships, accountability, taste, presence. Often a smaller number of more senior roles. Highly defensible but harder to enter.

The strategic question is which of the three you're becoming. Drift puts you in the second by default.

How to audit your weekly tasks

Use the AI career exposure audit and the career AI exposure worksheet for the full version. The fast version:

  1. List your top 10 weekly tasks.
  2. Tick each as routine, judgement-led, relationship-led, or physical.
  3. Count the rows. If routine is over 50% of your week, you're in the disruption path.

This isn't a verdict; it's a diagnostic. The follow-up action is to move the mix, not to panic.

What to do in the next 90 days

  1. Pick one routine task to automate or delegate to AI this week. Don't do all of them; pick one where you'll feel the time savings.
  2. Spend the saved time on the judgement-led part of your work. Strategic thinking, customer conversations, mentoring, planning, writing the brief that aligns 10 people.
  3. Read deeply about your industry's AI adoption. Not hype. Trade publications, internal memos, what your largest competitor is doing.
  4. Talk to one peer per week about how they're handling it. Honest conversations beat reading.
  5. Build one piece of public output — a written analysis, a talk, a project — that demonstrates judgement-led work.

What to do in the next 12 months

  1. Pick one durable, AI-resilient skill to deepen. See the high-income skills topic hub. 4-8 hours per week of deliberate practice.
  2. Build a 6-12 month financial buffer in cash. Not a luxury; insurance against forced career decisions during a restructure window.
  3. Map two adjacent roles you could move into. The roles your skill stack already partly fits. Reach out to one person in each within the next month.
  4. Audit your information diet. Cut one source that's reliably anxious without being informative.
  5. Have an honest conversation with your manager. Curiosity, not demand. Their read of your role's trajectory is data.

Common mistakes

  1. Confusing ‘AI-proof’ with ‘AI-resilient.’ The first doesn't exist.
  2. Treating disruption as a 5-year-away thing while your industry is already adjusting.
  3. Trying to learn five new skills at once.
  4. Ignoring AI tools because of taste or principle.
  5. Becoming the loudest doomer in the office. Nobody gets promoted for that.
  6. Burning bridges with the current employer before placing new bets.
  7. Not having a financial buffer.

Related

FAQ

Will I lose my job?
Honest answer: parts of your job are likely already substitutable; whether the whole role is restructured depends on the role, the company, and the next 24 months of capability deltas. The useful frame isn't ‘will I be replaced’ — it's ‘am I the version of this role that uses AI well, or the version AI replaces.’
How fast is this happening?
Faster than most company restructures handle gracefully, slower than the most dramatic predictions suggest. Specific roles (entry-level writing, basic coding, customer support, simple analysis) are seeing real substitution now in some companies. Most knowledge work is being augmented before being replaced. The 5-year window is the most uncertain.
What if I work in a regulated field — medicine, law, finance?
Slower disruption, but coming. Augmentation (AI drafting, AI search, AI summarisation) is already useful; full substitution is constrained by liability and regulation. The defensive move is to be the practitioner who uses AI well and remains the accountable human, not to ignore it.
Should I switch careers entirely?
Usually no, especially not in panic. Most of your skill stack transfers more than you think. Audit which sub-skills compound, which are AI-resilient, and which are commoditising fast; deepen the first two within your current career rather than starting over.
What about the ‘learn to code’ advice?
Coding is becoming more accessible, not less valuable, but the boundary is moving. The high-leverage skill in 2026 isn't typing code — it's specifying what to build, debugging integration, and judging when AI output is good enough to ship. Code as part of a broader skill stack.
How do I talk to my manager about this?
Privately, with curiosity, before you've made a decision. “I'm thinking about how AI affects my role. What's your read?” You learn more from one honest conversation than from a year of speculation. Most managers are also figuring it out — they'll usually meet honesty with honesty.