AutomationAtlas

What work tasks can AI assist with — and where in the world?

An open atlas of automation exposure across 124 countries, 23,851 tasks, and 6,275 skills. Based on task-level measurement, not aggregate predictions.

World View preview

Country hover

Hover or tap a country to inspect the current metric and move into the country surface.

Start here

Three ways in.

Pick a starting point. Each takes you to a different depth — broad comparison, single country, or guided narrative.

World map showing automation exposure by country
World View

Compare countries by automation exposure.

See all 124 countries on one map. Filter by income group or region, and switch between the country-aware benchmark and a simpler global comparison.

Open world view
Scatter plot of exposure versus GDP per capita
Country Profiles

Explore one country in depth.

Choose any country to see its full profile: average exposure level, top occupations, industries, skills most affected, and where it sits relative to its income peers.

Open country profiles
Case study preview
Case Studies

Read country diagnostics with context.

Guided narratives that explain what the numbers mean for a specific country — written for policy and general audiences, not just economists.

Browse case studies

What this atlas measures

Task exposure — not adoption, displacement, or wages.

We evaluated 23,851 real work tasks against what current AI systems can assist with. For each task, we asked: can an AI system plausibly handle or support this activity?

The result is a country-level map of which tasks are most amenable to AI assistance — and how that varies across income levels, regions, and trade-linked industries.

The atlas measures exposure at the task level. It does not predict whether automation will actually happen, or what effect it will have on employment or wages.

124
countries in the main panel
23,851
work tasks evaluated
6,275
skills mapped to tasks
4
exposure levels (1 = low, 4 = high)

Case studies

Country diagnostics in plain language.

Each case study packages the same public data into a readable country narrative — suitable for policymakers, researchers, and general readers.

Zambia case study
Case study

Zambia

Zambia sits above the current Sub-Saharan Africa median on average task exposure, but its profile still leans more toward augmentation than substitution. The strongest signals are concentrated in clerical, accounting, and digitally mediated service work, with specialized domain tools and software-heavy tasks doing most of the work.

Read Zambia
Kenya case study
Case study

Kenya

Kenya combines a higher average exposure level than Zambia with a similarly augmentation-heavy overall profile. The strongest signals again sit in administrative, clerical, and digitally mediated service work, making Kenya a natural comparative case for policy discussion in East Africa.

Read Kenya