evidoria

← Back to browse

Good practice Imported

AI at Work 2.0 — Malaysia's government-wide generative-AI productivity rollout for civil servants

Malaysia · Putrajaya · See the Malaysia profile · See the Putrajaya profile

Evidence: Descriptive / self-reported Top 66% 53/100 · Ask Evidence Copilot about this practice

Malaysia's Ministry of Digital and Google Cloud are rolling out Gemini-powered Workspace tools to up to 445,000 public officers; a World Bank/GovTech survey of 16,500+ civil servants across all 28 federal ministries found AI users saving an average 2.5 hours daily, though reporti

445,000 people
Civil servants targeted
2.5 hours/day
Average daily time saved (self-reported) (2025)
16,500+ civil servants
Survey respondents
26.1 %
Service-delivery data published externally (2025)
AI at Work 2.0 — Malaysia's government-wide generative-AI productivity rollout for civil servants

Details

Maturity
Scaling
Promoter
Ministry of Digital Malaysia, with Google Cloud; survey by World Bank & Public Service Department
Period
December 2024–ongoing (AI at Work 2.0 announced February 2025)
Keywords
generative AI, digital government, civil service productivity, cloud AI, workforce skills

Context

AI at Work 2.0 is a Malaysian government-wide rollout of Gemini-powered Google Workspace tools, led by the Ministry of Digital together with Google Cloud, giving roughly 445,000 civil servants across the federal government access to generative-AI features. It followed the December 2024 launch of Malaysia's National AI Office and was formally announced in February 2025.

Objectives

The program is intended to raise productivity among public officers through generative-AI assistance, build workforce skills in prompt-writing and AI tool use, and support the broader digital transformation of the civil service.

Activities

Civil servants get access to Gemini-enabled Google Workspace features including drafting and summarizing email and documents in Gmail and Docs, generating training videos in Google Vids, and building custom AI agents ('Gems') for repetitive tasks such as analysing citizen feedback, supported by hands-on prompt-writing workshops.

Results

A World Bank/GovTech survey of over 16,500 civil servants across all 28 federal ministries found that civil servants using AI tools reported saving an average of 2.5 hours per day, though adoption varied considerably by ministry — the Digital Ministry and Ministry of Investment led, while the Women, Family and Community Development Ministry and Finance Ministry lagged. The same reporting noted that only 26.1% of service-delivery data is published externally, versus a 35% OECD average.

Conclusions

Reported gains rest on self-reported survey data from a large but non-experimental sample, with no control group or randomized comparison; uneven ministry-level adoption and transparency gaps are flagged as open issues even as the rollout continues to scale government-wide.

Implementation
Indicative cost
High (€500k–€5M) — No public budget figures were found; enterprise Gemini/Google Workspace licensing plus training for ~445,000 civil servants across the entire federal government implies a substantial recurring cost, estimated as high in the absence of disclosed figures.
Time to results
Medium (1–3 years) — Launched alongside the National AI Office in December 2024 and formally announced as 'AI at Work 2.0' in February 2025, with rollout and adoption still ongoing as of late 2025 — a medium-length, still-unfolding implementation.
Staffing & skills
Ministry of Digital Malaysia (programme ownership), Google Cloud (Gemini/Workspace platform and technical delivery), Public Service Department (civil-service-wide rollout), World Bank / GovTech (independent survey of adoption and impact)

Conditions for success

  • Government-wide Google Workspace licensing and IT provisioning across 28 federal ministries
  • Hands-on prompt-writing training/workshops for civil servants
  • Ministry-level champions to drive adoption (Digital Ministry, Ministry of Investment led early)
  • Custom 'Gems' agents tailored to agency-specific repetitive tasks

Common failure modes

  • Adoption varies sharply by ministry, with Women, Family and Community Development and Finance Ministries lagging
  • Reported productivity gains are self-reported survey estimates, not independently verified time-and-motion data
  • Transparency gap flagged alongside the rollout: only 26.1% of service-delivery data published externally versus 35% OECD average

Where it fits

Governance type
national federal civil service
Scale
government-wide, ~445,000 civil servants across 28 ministries
Income level
upper-middle-income (Malaysia)

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful