evidoria

← Back to browse

Good practice Imported

Germany's Agentic AI Speeds Up Hydrogen Core-Network Permitting — 2026 GovTech Prize Winner

Germany · Hamburg · See the Germany profile

Evidence: Descriptive / self-reported Top 51% 60/100 · Ask Evidence Copilot about this practice

Germany's BMDS and BMWE built an agentic AI that drafts hydrogen-pipeline permit decisions in hours instead of weeks or months, with humans making final calls; live in Hamburg, it won 'Best Use of AI in Government Services' at the 2026 GovTech Prize.

Germany's Agentic AI Speeds Up Hydrogen Core-Network Permitting — 2026 GovTech Prize Winner

Details

Maturity
Pilot
Promoter
Federal Ministry for Digital and State Modernisation (BMDS) and Federal Ministry for Economic Affairs and Energy (BMWE)
Period
2025-11/2026-02
Region (NUTS)
DE60
Keywords
energy infrastructure, permitting, public administration, digital government

Context

Germany's Federal Ministry for Digital and State Modernisation (BMDS) and Federal Ministry for Economic Affairs and Energy (BMWE) built a nationwide digital application-and-approval platform, on the existing OZG-EfA "DiPlanung" infrastructure already used by several German states for spatial planning, to speed up permitting for the national hydrogen core network — previously, staff manually reviewed hundreds of pages of application documents per case, taking weeks to months.

Activities

The platform entered pilot operation in November 2025 for hydrogen core-network pipeline permits and was extended with an agentic-AI component that analyses submitted application documents, structures their content, checks whether required evidence is present, and drafts an informed decision proposal for caseworkers, while final decisions and legal/discretionary responsibility remain with the human employee. As of the award announcement, the AI agents are live in Hamburg for approving hydrogen core-network lines.

Results

The ministries report the review step drops from weeks or months to "a few hours" but have not published aggregate before/after processing-time statistics or the number of permits processed by the AI agent to date. The system's main third-party validation is the "Best Use of AI in Government Services" award at the GovTech Prize 2026 (World Governments Summit, Dubai, February 2026).

Conclusions

BMDS and BMWE plan to open-source the agents for reuse across federal, state and local governments, but that broader rollout has not yet happened.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
Federal Ministry for Digital and State Modernisation (BMDS) and Federal Ministry for Economic Affairs and Energy (BMWE) jointly built and run the platform, Human caseworkers review the AI's drafted decision proposals and retain final legal and discretionary responsibility

Conditions for success

  • Built on the existing OZG-EfA 'DiPlanung' infrastructure already used by several German states for spatial planning, rather than a new platform from scratch
  • AI agent limited to drafting/structuring and evidence-checking, with humans making the final call

Common failure modes

  • No aggregate before/after processing-time statistics or permit-processing volumes have been published
  • As of the award, live only in Hamburg — broader federal/state/local rollout has not yet happened despite plans to open-source the agents

Where it fits

Governance type
federal government (cross-ministry), coordinated with state/local permitting authorities
Scale
sub-national pilot (Hamburg) within a national programme
Income level
high-income

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