Research + Tool

Is your CS job designed
to be replaced by AI?

Two scoring systems. One conclusion. If a company can't talk about customer relationships in a job description, they're not building a CS function — they're building an automation workflow with a human placeholder.

8.2/10
BLS AI exposure — customer service reps
342
US occupations scored by Karpathy
0
BLS categories named "Customer Success"

Paste any job description. See if AI
empowers this role — or eliminates it.

Paste the full job posting or your internal job description. The scorer reads how the company talks about relationships vs. processes — and returns a displacement risk score, rationale, and the specific signals it found in the text.

What's your displacement risk?

Paste any CS job description — your current role or that job you're checking out.

No spam. Just the scorer + March 25 session details.

Reading between the lines...

Two ways to read
the same problem

Andrej Karpathy scored 342 BLS occupations by how much AI will reshape them. That's the macro lens — useful, but blunt. It can't tell you whether this CS role at this company is a real relationship job or a process workflow in disguise. That's what the second lens is for.

Lens 01 — Karpathy / BLS Data
The macro view:
occupation-level AI exposure

Scores every BLS occupation category from 0–10 by how much AI will reshape it. A score above 7 means the work product is fundamentally digital and the trajectory is steep. CS-adjacent BLS categories score high — because BLS still thinks CS is customer service.

Customer svc reps
8.2
Sales managers
6.5
IT project managers
6.7
Management analysts
7.2
Training specialists
6.0

Source: karpathy.ai/jobs — BLS Occupational Outlook Handbook, LLM-scored. Scores are approximate for CS-adjacent categories.

Lens 02 — Lincoln Scoring
The role view:
relationship language density

Reads the actual job description. How much does the company talk about human judgment, executive relationships, business outcomes, and navigating ambiguity — vs. processes, volume, tooling, and task execution? The ratio is the real displacement signal.

High relationship density
Safe
Mixed signals
Watch
Process/task dominant
At risk

Based on 15+ years of CS org design. See the scorer below.

"BLS doesn't have a category called Customer Success. It calls it customer service — and scores it accordingly."

That single categorization error is why macro AI-exposure data tells an incomplete story for CS professionals. The BLS is measuring the wrong thing.

The real question isn't whether the occupation is exposed. It's whether this specific role, as written, is doing real CS work — or executing an automation-friendly process that a company will eventually hand to an agent. The job description tells you. If you know how to read it.


March 25, 2026 — Live Session

Introduction to Agentic Workflows for Customer Success

AI isn't coming for CS. It's coming for the version of CS that was never really CS to begin with. This session covers what that means for your team, your role, and how to build a CS practice that AI makes stronger — not redundant.

Reserve your seat
Free live session with Lincoln Murphy. Limited spots.
DATE March 25, 2026
FMT Live online session + Q&A
FOR CS leaders, CSMs, CS ops
COST Free
Register now

Recording available to registered attendees only.