GitHub logo

How GitHub Cleared 100K Tickets with a Custom AI-Powered Zendesk System

GitHub partnered with Andela to deploy a high-impact team of technical experts to power AI-driven support, intelligent escalation, and code-level troubleshooting. By combining advanced tooling with expert talent, support evolved from a critical bottleneck into a strategic driver of customer success and growth.

Instructions from Copilot for transferring repository ownership outlining three steps: ensuring admin access, checking target account, and initiating transfer by clicking Transfer in the Danger Zone section; includes GitHub resource links and a support ticket message requesting guidance on transferring repo ownership.Instructions to transfer repository ownership including ensuring admin access, checking the target account, and initiating the transfer from the Settings 'Danger Zone'.

Snapshot

Duration
April 2024 – Ongoing
Team size
Portrait of a man with glasses, short curly hair, and a beard wearing a black jacket and brown shirt against a light background.Portrait of a woman with long braided hair resting her chin on her hands against a plain background.Portrait of a woman with curly dark hair, wearing glasses and a dark top, resting her chin on her hand.Smiling man with short hair and beard wearing a black jacket and black shirt against a light gray background.
+ 56 Andela specialists.
Technical domains
API Integration & Security
ML Classification
Analytics & Data Visualization
Microservices Architecture
Tech stack
Zendesk
Linux
Git
Bash
Ruby
ChatGPT
Splunk
Sentry
Kusto Query Language
100K

Tickets/yr cleared

4.61

CSAT Score

100%

SLAs consistently met

3X

Reduced Resolution Times

Table of contents
Business problem
Solution
Team curation
Technology
Impact

Business problem

A 100K-ticket backlog put trust, growth, and efficiency at risk

GitHub's support infrastructure—powered by strong internal tooling and Zendesk integrations—was built for reliability. But a backlog of 100,000 unresolved tickets began to impact efficiency and slowed ticket response times. To protect service quality and meet customer expectations, GitHub needed an immediate, scalable support intervention.

Trust and loyalty at risk

GitHub needed to maintain its reputation for reliability and meet customer expectations.

Revenue growth threatened

The high ticket backlog caused a potential impact to renewal rates.

Operational resource burn

An 18-month rotation policy drove costly turnover and constant retraining.
Closed GitHub support ticket titled 'I need guidance to transfer old repo ownership' by Roberto Montero, showing a comment requesting steps to transfer repository ownership.

THE SOLUTION

An AI custom-built Zendesk system, powered by Andela’s AI engineers

Serving millions of developers and organizations daily, GitHub built a robust customer support system centered around Zendesk, deeply integrated with its proprietary internal tools. This framework was designed to ensure rapid, reliable service at scale, combining automation with human expertise:

AI-Driven ticket management

ML models and scripts automate ticket triage at scale.

API-aware self-resolution

An API-integrated knowledge base enables developer self-service.

Integrated tools for rapid resolutions

Custom scripts connect Zendesk and GitHub for code-level issue tracing.

Data-driven optimization

KQL and Splunk insights improve workflows and SLA performance.
GitHub Support page displaying a form titled 'Get help with GitHub' where a user is filling out account details, subject, and message requesting guidance to transfer repository ownership.

Handing off the baton to Andela's engineer support team

To clear the backlog without disrupting operations, GitHub partnered with Andela to deploy a fully managed support team of technical engineers—running, optimizing, and evolving its custom Zendesk framework and broader support architecture.

Expert, scalable support teams

Andela deployed 56 engineers to resolve tickets and optimize support.

Seamless integration

Teams integrated into GitHub’s existing tools without disrupting workflows.

Framework optimization

Engineers streamlined tagging and queue workflows to improve speed and accuracy.

End-to-End ownership

Support engineers managed advanced CI/CD troubleshooting through secure data remediation.
Andela website content image
We don’t develop or customize Zendesk internally; instead, we excel at managing, categorizing, and rapidly resolving large volumes of tickets. Our approach relied on drive, technical knowledge, and a relentless focus on clearing a backlog of 100,000 tickets in under a year.
Oluwaseun Ayodeji
Team Lead, Technical Support
Andela website content image

the team

A support system built for developers, powered by developers

Andela deployed a 56-person support engineering team across four specialized squads—Account Security & Billing, Compliance, Enterprise, and Technical Support—each led by a Team Lead. Embedded in GitHub’s stack (Zendesk, Linux, KQL, Splunk, AI tools), they resolved tickets while optimizing workflows and improving internal systems without new integrations.

Technical Support Leads
Portrait of Andela leadership team memberPortrait of Andela leadership team member
Jeffrey Ogah
Account Security and Billing
Click for more
Jeffrey Ogah
Account Security and Billing
Focused on less technical issues, such as billing inquiries, account lock-outs, and 2FA support.
Focus
Auth Systems
Payment APIs
Security Protocols
Portrait of Andela leadership team memberPortrait of Andela leadership team member
Anastasia Rotich
Compliance
Click for more
Anastasia Rotich
Compliance
Assists users with account takeovers, spam/flagged issues, ownership disputes, username requests, and trade restrictions.
Focus
Auth Systems
Compliance Automation
Data Protection
Portrait of Andela leadership team memberPortrait of Andela leadership team member
Maureen Oki
Enterprise
Click for more
Maureen Oki
Enterprise
Managed deployment-related challenges, often requiring Linux expertise for diagnosing server-side or networking errors. Focused on offloading the influx of tickets related to GitHub Copilot and other high-demand services.
Focus
Large-Scale Deployment
Enterprise Integration
SSO
Portrait of Andela leadership team memberPortrait of Andela leadership team member
Oluwaseun Ayodeji
Technical Support
Click for more
Oluwaseun Ayodeji
Technical Support
Handled mid-level tasks like debugging technical errors and assisting users with config-related issues.
Focus
Git Internals
API Troubleshooting
CI/CD Pipeline

THE TECH

Driving resolution through AI, automation, and deep system integration

GitHub’s support stack was built for developer-grade speed and complexity. Its deeply integrated Zendesk system—enhanced by automation, Copilot, and AI—enabled intelligent triage, contextual routing, and high-velocity resolution at scale.

KQL

Queried customer data and logs to isolate issues quickly.

Splunk

Analyzed logs to detect authentication and server-side anomalies.

Sentry

Monitored real-time errors to support faster debugging.

Linux and Git

Reproduced customer issues in self-hosted and Linux environments.
Andela website content image
Our process requires deep knowledge of Git, Linux, and networking. To reproduce customer issues, we’ll write scripts, analyze logs, and understand the full software development lifecycle, all to deliver precise solutions and streamline customer support at scale
Oluwaseun Ayodeji
Team Lead, Technical Support
Andela website content image

Zendesk Implementation

Andela’s engineers operated directly within GitHub’s Zendesk environment to resolve issues at scale. Trained on internal systems and workflows, they used historical data, automation, and AI-powered triage to identify patterns, prioritize accurately, and continuously improve queue performance.

Centralized ticket routing

All inbound tickets were funneled into Zendesk for structured triage and SLA-based management.

Categorization and collaboration

Engineers labeled tickets, added notes, and routed issues to the right GitHub teams.

Analytics and reporting

Metadata was tracked for trends, enabling proactive insights and performance optimization.

System integration

Custom Zendesk flows linked tickets to GitHub repos for faster debugging and resolution.

THE impact

The ROI of turning technical support into a strategic asset

As GitHub faced mounting operational strain, it turned to Andela for fast, scalable expertise. By embedding into GitHub’s support stack, resolving 100,000 tickets, and maintaining a 4.6+ CSAT, the partnership eliminated bottlenecks and positioned support for long-term scale.

Resolved customer tickets
100K

Cleared all support tickets within one year.

Customer Satisfaction Score
4.61

Maintained a consistent CSAT score.

GitHub Trained Experts
47

With additional certifications in CoPilot, Admin, Actions, and Advanced Security.

Service Level Achievement
100%

Consistently met all SLAs for one year.

Reduced Wait Times
3X

Dramatically improved customer experience through faster resolution.

Bypassed Microsoft's Rule
18 Mon

Removed contractor 18 months limitation to maintain knowledge and expertise.

Explore more customer stories

Andela partners with F250s to build against AI business outcomes.

80%
Database uptime boost

GoPuff Boosts Uptime with Azure Flex & HA Postgres

33%
Faster project delivery

SoFi Powers AI-Driven Risk Intelligence

Partner With Andela 

Whether you need individual technologists or fully managed teams, we can transform your support operations from a cost center into a strategic advantage that enhances your brand and offering.

Book a discovery call
Profile card of Samantha C., Senior Data Analyst based in Brazil, GMT+2 with 4 hours overlap, earning $6,500 - $8,500 per month.Profile card of Samantha C., Senior Data Analyst from Brazil, GMT+2 timezone, with 4 hours overlap and salary range $6,500 - 8,500 per month.

Subscribe for insights on building and scaling AI systems.

Decorative graphic element
Field required
Thank you, we received your contact information!