LinkedIn Automates Software Licenses

Managing software licenses is a sprawling, often inefficient, process for large organizations. At LinkedIn, the challenge of granting and revoking access for thousands of employees across hundreds of applications led to significant productivity drains and overlooked costs. Licenses went unused while employees waited for approvals, and accounts lingered long after employees departed. To combat this, LinkedIn engineered its own solution: a centralized platform for automating software license provisioning and reclamation, dubbed SLLM.

This initiative, detailed on the LinkedIn Engineering blog, aims to streamline the complex enterprise software license lifecycle. The vision was to separate strategic procurement from day-to-day operational efficiency, focusing SLLM specifically on the automated provisioning and reclaiming of licenses.

Core Principles for SLLM

The development of SLLM was guided by four core principles: automation first, a user-centric experience, data-driven intelligence, and leveraging existing systems over building from scratch.

This guiding philosophy set the stage for a three-phase development journey.

Phase 1: Intelligent License Reclamation

The initial focus was on reclaiming underutilized licenses. This involved identifying two key user categories: terminated employees, cross-referenced against active directory records, and inactive users flagged after a configurable period of non-usage. Crucially, the system incorporates safeguards like extended leave protection, legal hold integrations, and exclusion lists to prevent erroneous reclamations.

Phase 2: Streamlined License Provisioning

With reclamation in place, the team turned to provisioning. A dual-track approach caters to diverse application needs. Request-based provisioning handles applications requiring user initiation, either through Active Directory group integration for seamless SSO or via custom API calls for more complex scenarios. A pre-approval mechanism allows for automatic access for specific user cohorts.

The most impactful innovation was zero-touch provisioning via app bundles. These bundles automatically assign licenses based on employee attributes like department or job title. This system relies on bundle definitions, a validation engine linked to the identity management platform, and an execution layer that queues and processes requests in batches to prevent errors.

Phase 3: Enriched Employee Experience

The final phase concentrated on user interfaces. A comprehensive dashboard provides employees and administrators visibility into current licenses, usage statistics, and reclamation timelines. Notifications alert users to potential reclamations, offering a chance to justify continued need.

Integration with the IT service portal also enables natural language requests for new software, abstracting technical complexities for end-users.

Technical Implementation and Impact

Key technical decisions included a configuration-driven design for flexibility, batch processing with smart aggregation to handle scale, and multi-model integration to support various application architectures. The impact has been substantial, with SLLM now covering 90% of software provisioning requests and onboarding 53 critical applications. Zero-touch provisioning has processed around 10,000 requests in four months, significantly speeding up new joiner onboarding.

Enhanced visibility into software usage aids optimization and contract renewals, while automated reclamation bolsters security and compliance by promptly revoking access for departed employees. This robust system adheres to vendor licensing models and supports internal audits.

Lessons Learned

Building SLLM underscored the importance of cross-team collaboration, particularly between app owners and operations. Data quality was identified as foundational, requiring significant investment in validation and reconciliation.

Safeguards were prioritized over raw speed to prevent potential system-wide issues. A progressive rollout strategy allowed for incremental learning and stakeholder buy-in.

The team also highlighted the value of structured onboarding processes for new applications, ensuring quality and trust. The platform continues to evolve, with plans for advanced analytics and enhanced recommendation systems.