World Bank Group Development Data Group
For National Statistical Organizations

World Bank Group AI-Readiness Assessment Framework

Assesses both the institution and the data products and services it provides.

World Bank Group Development Data Group
Development Data GroupOffice of the WBG Chief Statistician
Illustration of AI-ready official statistics: a glowing brain over a national statistics building, surrounded by icons for legal, skills, technical environment, funding, partnerships, API access, agentic access and trusted provenance
2pillars
12dimensions
69questions, plus a gateway question
4maturity levels, A to D
Value added

What the WBG assessment adds to other assessments

Existing assessments, such as SPEED, give a broad institutional diagnostic. This assessment adds depth on data products and ties results to the guidance of the Committee for the Coordination of Statistical Activities (CCSA).

SPEED
28
indicators across 5 pillars
WBG assessment
69
questions across 12 dimensions, plus a gateway question

Bars are drawn to the same scale. The assessment is longer and asks more detailed questions.

1

Data product depth

Tests whether data products and services are ready for AI: catalog, APIs, agentic access, licensing and quality. SPEED covers this side with a few indicators.

2

Readiness and use

Scores whether the conditions for AI are in place, and separately how far AI is already used. Seeing both shows where use is ahead of the foundations.

3

Tied to CCSA guidance

Each question maps to the CCSA recommendations for NSOs, so results point to internationally endorsed guidance. See the annex.

Works for NSOs at any starting pointAlso for central banks and ministries that produce official statisticsOpen license, CC BY 4.0Web form scores and reports automatically
The framework at a glance

One framework. Two pillars. Twelve dimensions.

Select any dimension to read what it assesses. Use the arrow keys to move through them.

Pillar I · AI for Data

Institutional Readiness for AI Adoption

Can the NSO use AI responsibly?

Pillar II · Data for AI

Readiness of Data Products and Services for AI

Can AI use official statistics reliably?

Full text from the framework document

Related CCSA recommendations
Why two pillars

The two pillars point to different kinds of action

An organization can score high on one pillar and low on the other, and each case calls for different action. Pick which pillar lags.

Pillar I · AI for Data

Institutional capacity to act

Typically requiresStrategic and organizational change
Calls for investment inGovernance and capacity
When this pillar lagsData products can be open and API-accessible while the institution still lacks the foundations to adopt AI responsibly.
Pillar II · Data for AI

Data products ready to serve

Typically requiresTechnical and operational change
Calls for investment inData product modernization
When this pillar lagsGovernance, skills and infrastructure can be strong while data products are not yet modernized for machine consumption.
Assessment design

Questions become evidence, scores, and an actionable profile

The instrument supports self-assessment or facilitated assessment across NSOs at any starting point.

Readiness

Conditions that enable safe and effective AI

Adoption & maturity

Extent and sophistication of AI use

What the assessment produces

A readiness profile that can move directly into action planning

Click a cell to set a level for each dimension. The panel shows where gaps sit and how they cluster. It starts with the illustrative profile from the deck.

Click the same cell again to clear a dimension.

Pillar view

Pillar I · AI for Data
Pillar II · Data for AI

Priority gaps

    Sandbox for exploring the format. Here the bars fill from A to D, as in the deck's illustrative profile. Expert review and institutional judgment remain essential.

    Priority gaps

    Where readiness is constrained

    Sequenced actions

    What to address first

    Partnership needs

    Where external support adds value

    Complementary value

    Assessment should lead to a country-owned modernization pathway

    The framework complements partner assessments by adding an explicit lens on AI-ready official statistics.

    What to strengthen

    Priority institutional and data-service gaps

    What to build

    A realistic sequence aligned with national capacity

    Where to partner

    Shared tools, standards, financing and expertise

    Annex · Tied to CCSA guidance

    39 recommendations for national statistical organizations

    From AI-Ready Official Statistics: Opportunities, Challenges, and Recommendations, a background document for the 57th session of the UN Statistical Commission (March 2026). The recommendations for NSOs run from R.1 to R.39 in five areas. Open the paper (PDF)

    Summaries are condensed from the paper; see the PDF for the exact wording. The links from each dimension above are an indicative mapping made for this page. The assessment instrument holds the official question-level mapping.