Last week we talked about mapping your dependencies (identifying the single points of failure that make you vulnerable to organizational decisions you don't control). This week we're going deeper: assessing which parts of your actual job are defensible against automation, and which aren't.
Most people don't realize which of their capabilities are at risk until they're already being replaced. By then, your options narrow considerably. The goal here isn't to panic you. It's to help you see your position clearly enough to make strategic decisions about where to invest your development time.
Here's a framework to do exactly that.
The Three-Axis Assessment
To understand which parts of your role are genuinely defensible, you need to evaluate them across three dimensions:
1. Automation Difficulty (Low / Medium / High)
How hard would it be to automate this capability with current or near-term AI?
Low difficulty: Pattern-based work with structured inputs and predictable outputs. If the task follows clear rules or relies on matching historical examples, it's automatable. Think: scheduling, basic data analysis, templated writing, initial CV screening.
Medium difficulty: Work that requires some contextual judgment but operates within defined boundaries. Mixed structured and unstructured inputs. Examples: performance review drafting, training needs analysis, straightforward stakeholder management.
High difficulty: Deeply contextual judgment, political navigation, work that requires reading unspoken dynamics or synthesizing genuinely novel situations. Things like: sensitive employee relations, strategic workforce planning, managing competing executive agendas.
The question isn't "could AI do this?" It's "would it be cheaper and easier to automate this than to keep paying me?"
2. Value Extraction (Low / Medium / High)
How much cost reduction or efficiency gain does automating this create for the organization?
Low value extraction: Automating it saves relatively little money or doesn't significantly speed up critical processes. Niche capabilities, small-scale work, things that don't happen often enough to justify automation investment.
Medium value extraction: Clear cost savings or efficiency gains, but not transformational. Worthwhile to automate eventually, but not the first priority.
High value extraction: Automating this capability saves significant money or dramatically accelerates important workflows. These are the tasks finance teams notice. High-volume work, expensive specialist work that can be commoditized, bottlenecks in critical processes.
This isn't about your value to the organization. It's about the financial case for replacing you.
3. Strategic Importance (Low / Medium / High)
How central is this capability to the organization's competitive advantage or strategic priorities?
Low importance: Necessary but not differentiating. Standard operational work that every similar organization does roughly the same way. If outsourcing it wouldn't raise eyebrows, it's probably low strategic importance.
Medium importance: Contributes to organizational effectiveness but isn't core to what makes the company unique. Competence here matters, but it's not what the C-suite talks about when they describe their strategy.
High importance: Core to organizational differentiation, leadership priorities, or competitive positioning. Work that connects directly to strategic goals. Capabilities that, if done poorly, would genuinely threaten the business.
The key insight: strategic importance can protect you even when automation difficulty is low. If leadership believes this capability is core, they're less likely to commoditize it.
The Four Outcomes
When you plot your capabilities across these three dimensions, patterns emerge. Here's what they mean:
DEFEND (high across all dimensions): These are your moat. Difficult to automate, expensive to replace, strategically important. Double down here. Make these capabilities more visible, more integrated into critical workflows, harder to separate from you personally.
DEVELOP (mixed scoring): These capabilities have potential but need strengthening. Maybe they're strategically important but too easy to automate (add complexity, deepen expertise). Or they're hard to automate but low strategic importance (connect them to leadership priorities). Identify which dimension to strengthen.
DIVERSIFY (low automation difficulty, regardless of other scores): These are getting automated eventually. The timeline might be long, but the direction is clear. Don't abandon them (you still need to do your job) but start building capabilities in other quadrants. You're hedging, not fleeing.
DOCUMENT (high automation risk, high value extraction): These are actively being evaluated for replacement. If you're in this quadrant, focus on gathering proof of impact now. Build your evidence portfolio. Make the case for why human judgment matters here. And quietly start moving capabilities toward other quadrants.
How to Actually Run This Audit
Here's the process:
Step 1: Break down your job into discrete capabilities
Don't just list job description bullets. Identify actual capabilities you use:
- What do you do that requires genuine skill or judgment?
- What tasks do you perform repeatedly?
- What do stakeholders actually come to you for?
Be specific. "Stakeholder management" is too broad. "Managing conflicting priorities between finance and operations during budget season" is a capability you can assess.
Step 2: Score each capability across the three dimensions
Use your honest judgment:
- Could current AI handle 80% of this task adequately? (Automation Difficulty)
- Would automating this save meaningful money or time? (Value Extraction)
- Is this connected to what leadership cares about? (Strategic Importance)
Don't overthink it. You're looking for patterns, not precision.
Step 3: Plot them and identify clusters
Where are most of your capabilities landing?
- Heavily in DEFEND? You're well-positioned.
- Mostly in DEVELOP? You've got work to do but you have time.
- Clustered in DIVERSIFY or DOCUMENT? You need to move deliberately.
Step 4: Choose one action per quadrant
Don't try to fix everything. Pick:
- One DEFEND capability to make more visible
- One DEVELOP capability to strengthen
- One DIVERSIFY capability to start building an alternative for
- One DOCUMENT capability to gather evidence on
That's your roadmap for the next quarter.
A Real Example: Career Coaching
Let me show you what this looks like with something I've been thinking about a lot: one-on-one career coaching and development conversations.
This has been a core L&D capability for decades. Managers do it informally, HR professionals do it structurally, external coaches do it commercially. It's relationship-based, contextual, requires reading the person in front of you and synthesizing their specific situation.
Here's how I'd score it:
Automation Difficulty: Medium (and dropping fast)
Two years ago, I'd have said High. Career coaching requires understanding someone's context, values, constraints, and aspirations, then helping them synthesize options they hadn't seen.
But AI is now doing this at scale, and it's "good enough" for most people. ChatGPT can conduct a structured career conversation, ask clarifying questions, identify skill gaps, suggest development pathways. It's not as good as an expert human coach, but it's vastly better than no coaching at all, and most people never get expert human coaching anyway.
The baseline has shifted. What used to require expensive one-on-one time is now available 24/7 for free.
Value Extraction: High
Organizations spend significant money on career development: internal L&D headcount, external coaching contracts, manager time in development conversations. If AI can handle 70% of career coaching adequately, that's a massive cost reduction.
And it scales instantly. One AI assistant can support thousands of employees simultaneously. No calendaring, no capacity constraints, no geographical limits.
Strategic Importance: Medium
Here's where it gets interesting. Most organizations claim career development is strategically important ("our people are our competitive advantage"), but few actually differentiate on it. It's table stakes, not advantage.
Unless you're a company where talent development is genuinely core to your business model, career coaching sits at Medium. Important enough that you need to do it, not important enough that how you do it matters strategically.
Outcome: DEVELOP (with urgency)
The automation difficulty is falling. The value extraction case is clear. The strategic importance isn't high enough to protect it.
So what would I do if this were my primary capability?
- Increase automation difficulty: Move up-market. Focus on the cases AI can't handle well: complex political situations, sensitive performance issues, executive transitions, coaching through organizational upheaval. Specialize in the genuinely hard conversations.
- Increase strategic importance: Connect career development explicitly to business priorities. Don't just help people grow. Help them grow in ways that advance strategic goals. Make your coaching about organizational capability-building, not just individual development.
- Document impact: Start gathering proof that human coaching creates outcomes AI coaching doesn't. Track career moves, promotion rates, retention of coached employees. Build the business case for why this still needs you.
That's the strategic response to seeing your capability position clearly.
What to Do With Your Results
Once you've assessed your capabilities, here's the action guidance for each quadrant:
DEFEND:
- Make these capabilities more visible to leadership
- Integrate them deeper into critical workflows
- Mentor others in these areas (increases your strategic importance)
- Document your methods so they're associated with you personally
DEVELOP:
- Identify which dimension is weakest and strengthen it deliberately
- Connect your work to strategic priorities more explicitly
- Build relationships with stakeholders who can validate importance
- Add complexity or depth that makes automation harder
DIVERSIFY:
- Don't abandon these (you still need them for your current role)
- Start building capabilities in DEFEND/DEVELOP quadrants
- Create proof of work while you still can
- Expand your skill range strategically, not randomly
DOCUMENT:
- Build your evidence portfolio now
- Make the case for human judgment in this domain
- Gather testimonials, impact metrics, before/after examples
- Prepare for the conversation about whether this should be automated
The goal isn't to become automation-proof. Nothing is. The goal is to see your position clearly enough to make strategic decisions about where to invest your limited development time and energy.
The Adaptive Assessment
Here's the thing: running this audit manually is valuable, but it's also incomplete. You're working with your own biases about what's automatable, what's strategically important, and what's actually at risk.
I've built something that takes this further. It's an adaptive assessment that evaluates your specific role against real automation trends and organizational priorities, then generates a personalized development plan.
You can take it here: lead-assessment.thehumanco.org
It asks questions I can't cover in an article, adapts based on your answers, and benchmarks your position against what's actually happening in the market. Then it gives you prioritized actions: not generic advice, but specific moves based on your assessment.
This week: Pick three capabilities from your job. Score them honestly across the three dimensions. See where they land. Then choose one action from the guidance above.
You can't defend everything. But you can defend the right things.