How data scientists are reshaped as AGI capability advances.

Roughly 95% of the work in Data Scientists is information-shaped — already within reach of AI delivery. The question here is not whether it shifts, but which tasks go first and who staffs the residual.
Why: All 30 tools used by this occupation fall into UNSPSC segment 43 (IT/software), establishing a very strong digital baseline. This is corroborated by top work activities such as 'Analyzing Data or Information' (4.73) and 'Working with Computers' (4.68), alongside a work context dominated by 'E-Mail' (4.95) and 'Spend Time Sitting' (4.77), firmly placing this in the pure digital knowledge work band.
grounded in the economy graph · digital scalar 0.95 · digital
Read as an executable program — the work decomposed into Code, Generative, Agentic, and Human.
The work of Data Scientists engages 41 activities — the executable steps that, decomposed, reveal what becomes Code, what stays Human.
+29 more via engagesIn
Data Scientists involves 41 work activities — the generalized motions beneath the role, each scored against the AI-deliverability frontier.
+29 more via involvesActivity
Which of this work becomes digital labor — performed under typed authority, promoted to autonomy on track record.
Data Scientists performs 54 tasks on the graph — the atomic work units that become the job description for a digital employee, promoted to autonomy on track record.
+42 more via performs
Data Scientists is typically employed by 256 company types — the demand side that decides which of this role's tasks get handed to agents, and on what authority.
+244 more via typicallyEmploys
Data Scientists is employed across 243 settings — the places where this role's work is done, and where digital employees first sit beside the humans.
+231 more via employs
Data Scientists is linked from 2 entities via `performedBy` — a real edge on the economy graph, surfaced here so the claim stays grounded in data rather than assertion.
The software here going agent-consumable — where the API, not the UI, becomes the way the work gets done.
Data Scientists uses 9 tools today. As each gains an agent-consumable surface (API / MCP / SDK), the human UI stops being the only way in — and the work routes straight to an agent.
Data Scientists relies on 270 products. The headless dimension of each — whether an agent can call it without a screen — is what decides how much of this work goes hands-free.
+258 more via uses
The software Data Scientists reaches for already exposes 12 agent-callable actions (via uses → exposedBy) — typed surfaces an agent invokes directly, no human screen in the loop. The work routes to the API, not the UI.
Node-intrinsic problems read straight off the graph (exposesProblem) — the evergreen wedges a builder could take into this space.
+4 more problems on the graph
No articles yet for this entity.
No capability events for this entity yet.