Normalized claim
Adoption growth: 18 x increase
Since launch in Q1 2026, the agent’s adoption outpaced initial assumptions by 18x in that quarter.
GCPEvidence: Medium50/100
Delivery Hero’s Herogen autonomous software delivery agent using Agent Platform
Use case typeCoding assistantsUpdated Jun 30, 2026
Delivery Hero built Herogen, an autonomous software delivery agent that picks up Jira tasks, writes code, runs tests, and reviews itself. The system uses Claude for coding and Gemini within a council of agents for code review and security checks, with Google Cloud Agent Platform and LiteLLM as the enterprise foundation.
- Organization
- Delivery Hero
- Industry
- Consumer & Food
- Location
- Germany
- Published
- June 2026
Reported outcomes
18x
adoption growthAdoption & scale
170 PRs/daypull requests per day+10%total PR volume+85%production success rate
Strategic outcomes
Other strategic outcomeReduced engineer task time from a week to minutesCost efficiencyLowered developer toil across a decentralized engineering organizationSpeed & agilityPreserved local autonomy while standardizing AI infrastructure
Catalog median for adoption & scale deployments: +250% across 8 reported metrics. Compare benchmarks →
Why do we believe this?Outcome claims, sources, and evidence checks
Normalized claim
Pull requests per day: 170 PRs/day increase
Herogen’s autonomous agent creates and merges more than 170 pull requests every day
Normalized claim
Total PR volume: 10% increase
already accounts for over 10% of the company's total PR volume.
Normalized claim
Production success rate: 85% increase
Currently, Herogen boasts an 85% success rate for tickets merged to production
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- Delivery Hero
- Provider
- GCP
- Maturity
- Unknown
- Linked source
- Google Cloud Customer Story
No explicit deployment-stage evidence found.
Customer identity supportedSource describes one deploymentMaturity evidence evaluated
Primary read
Use case focus
Showing 3 of 3
- 1Coding assistants
- 2Workflow automation
- 3Agent orchestration
- Engineering velocity and developer toil across a decentralized, multi-subsidiary engineering organization.
- Needed a unified AI strategy that scales globally without losing local autonomy.
- Built Herogen as an autonomous software delivery agent.
- Used Google Cloud Agent Platform with LiteLLM to provide access to multiple models from engineers' IDEs.
- Used Claude as the primary coding model and Gemini for code review and security checks in a council-of-agents setup.
- Adoption grew 18x vs initial assumptions in Q1 2026.
- Herogen creates and merges hundreds of pull requests per day and accounts for 10%+ of total PR volume.
- Tasks that took a week for engineers are completed in minutes.
- Autonomous agents currently maintain an 85% success rate for tickets merged directly into production.
Architecture
Delivery Hero built Herogen, an autonomous software delivery agent on Google Cloud Agent Platform. The system uses LiteLLM as an LLM proxy so engineers can access models from their IDEs, with Claude handling coding tasks and Gemini used in a council of agents for code review and security checks. Herogen picks up Jira tasks, writes code, runs tests, and reviews itself.
Sources & evidence1
Evidence: Medium50/100Evidence strength
- Customer explicitly identified
- Primary source available
- Quantified outcome available
- Technical implementation details available
Type: Customer StoryPublished: Jun 30, 2026Publisher: Google CloudEvidence: PrimaryConfidence: High
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
AI patterns
Browse the catalog
Was this useful?
Community
Comments
No published comments yet.