Who Should Own Your Company's AI Layer? Insights from Glean’s CEO

Artificial Intelligence (AI) is no longer just a fancy chatbot answering FAQs. Modern enterprise AI systems actually perform tasks across organizations—streamlining work, automating decisions, and integrating deeply into business processes. In this evolving landscape, a critical question arises: who owns your company’s AI layer?

This isn’t a trivial issue. The AI layer acts as the backbone of an organization’s intelligent operations. Unlike isolated AI tools, this layer connects, processes, and enables the entire enterprise to harness AI effectively. Glean’s CEO recently shed light on this rarely discussed but fundamental challenge.

Why Is Ownership of the AI Layer a Complex Problem?

Ownership of the AI layer influences how organizations control, maintain, and evolve their AI capabilities. Often mistaken for just a technology decision, the answer touches on organizational structure, responsibility, and strategic priorities.

The problem starts because AI is shifting rapidly. Initially, AI was synonymous with chatbots or answering queries—relatively simple use cases. Today, it’s no longer just answering questions but actively doing work—such as generating reports, driving customer engagement, or automating parts of the supply chain.

In this context, multiple stakeholders within companies may claim ownership:

• IT Teams who manage infrastructure and data platforms.
• Product Teams who integrate AI into customer-facing features.
• Business Units needing AI tailored to their workflows.
• Data Science and AI Teams building models and AI pipelines.

Each party has a valid point, creating potential confusion and inefficiency. Without clear ownership, implementations risk becoming fragmented or stalled.

How Does Enterprise AI Differ From Earlier AI Forms?

To assess who should own the AI layer, one must understand what the AI layer entails today. It’s a complex system composed of several parts:

• Data ingestion and preparation: Collecting and cleaning data, a foundational step for AI accuracy.
• Model development and deployment: Creating algorithms that turn data into actionable outputs.
• Integration and orchestration: Embedding AI workflows into business processes.
• Monitoring and governance: Ensuring AI outputs are accurate, fair, and compliant.

This layered architecture is a far cry from simple Q&A systems. The AI layer must be both robust and flexible to adapt as business needs change.

Who Should Own the AI Layer, and Why?

Glean’s CEO argues that the AI layer ownership cannot be pigeonholed into a single department. Instead, it must be a collaborative model integrating technology and business leadership. Here’s why:

• Centralized expertise: AI specialists maintain models and infrastructure, ensuring technical soundness.
• Business context: Product and business teams steer AI to solve real-world problems aligned with company goals.
• Data stewardship: Ensuring trustworthy and compliant data flow requires oversight often crossing silos.

Ownership becomes a shared responsibility orchestrated through clear governance frameworks rather than a one-off project handoff.

When Does a Collaborative AI Layer Ownership Model Work Best?

This model thrives in organizations ready to treat AI as strategic technology—not just a feature or pilot. It requires:

• Strong leadership alignment across departments.
• Defined roles and responsibilities documented in governance policies.
• Investment in common AI platforms that enable reuse and scalability.
• Periodic review mechanisms to adapt AI as business needs evolve.

Ignoring these points often leads to duplicated effort, fragmented AI applications, or models that fail to deliver meaningful business value.

What Are the Trade-Offs of Different Ownership Approaches?

Some companies attempt to silo AI layer ownership within IT or data science teams, believing this will standardize output. However, this risks missing important business nuances, leading to tools that are technically sound but lack adoption.

Conversely, delegating ownership solely to business units leads to siloed efforts, incompatible AI systems, and inefficient maintenance.

Therefore, balancing centralized platform expertise with decentralized business adoption is challenging but necessary. Companies must be cautious not to oversimplify AI governance.

How Can Companies Implement Effective AI Layer Ownership?

To create an effective shared ownership, organizations can follow these steps:

• Set up a cross-functional AI governance board including IT, business, and AI experts.
• Develop standard operating procedures for model development, deployment, and monitoring.
• Choose AI platforms that support collaborative workflows and integration standards.
• Establish key performance indicators (KPIs) to measure AI impact and compliance.

This method brings transparency and agility, reducing risk as AI scales in complexity.

Real-World Example: Glean’s Experience

Glean’s journey shows the pitfalls of fragmented AI efforts. Early implementations suffered from unclear ownership—models built without business input led to underused tools. By shifting to shared ownership:

• Teams improved collaboration and faster iteration cycles.
• They aligned AI development with actual user needs.
• The company enhanced AI system reliability and governance.

These changes elevated AI from just another project to a core organizational capability.

How to Decide Who Should Own Your AI Layer?

To choose the right model for your company’s AI layer ownership, use this checklist:

1. Assess current AI maturity and organizational readiness.
2. Identify all stakeholders involved in AI development and usage.
3. Map overlapping roles, responsibilities, and gaps.
4. Agree on collaboration mechanisms and governance structures.
5. Commit to investing in platforms and skills to support shared ownership.

Investing time in this process can prevent costly failures and unlock the true potential of enterprise AI.

In summary: Owning the AI layer is not a technical detail—it reflects how an organization operates across silos, handles innovation, and delivers sustained business impact. Embracing shared ownership with clear governance can transform AI from a fragmented toolkit into a strategic asset.