What Is Unity Gateway?
As organisations introduce more AI models, applications and autonomous agents, managing how these technologies are accessed, used and governed is becoming increasingly important. Teams need the flexibility to adopt new AI capabilities while maintaining control over security, costs and performance.
Unity Gateway is Databricks solution for governing AI interactions through a central control plane. Built on Unity Catalog, it helps organisations manage access to AI models, agents, Model Context Protocol (MCP) servers and tools, while providing visibility and the controls on usage, costs and activity.
But what does Unity Gateway do, and how can it help organisations scale their use of AI?
What is Unity Gateway?
Unity Gateway is a centralised governance and management layer for AI services within the Databricks platform. It extends the governance capabilities of Unity Catalog beyond data and AI assets to the interactions between models, agents and the tools they use.
Rather than managing AI access, security policies and usage separately across multiple services, (Think Claude, CoPilot and Codex, what ever a users preference is) organisations can use Unity Gateway to apply consistent controls from one place.
Key capabilities include:
- AI access and governance: Control which users and applications can access models, agents and tools through identity-aware permissions and policies.
- AI cost management: Monitor consumption, attribute spend to users and teams, and set budgets, rate limits and spending caps.
- Model routing and traffic management: Manage requests across supported models and providers, with options for traffic distribution and failover.
- Security and guardrails: Apply policies to AI requests and responses, including controls to help protect sensitive information.
- Monitoring and observability: Track AI usage, token consumption, latency and activity to support operational oversight and investigation.
Together, these capabilities give organisations greater visibility and control over how AI is used across their technology estate.
Why is Unity Gateway important for enterprise AI?
As AI adoption expands, managing individual models and applications becomes more complex. Different teams may use different providers, deploy their own agents and connect AI systems to business tools and data. Without consistent governance, it can become difficult to understand usage, manage costs and maintain appropriate access controls.
Unity Gateway addresses several of these challenges.
1. Managing AI costs
AI services can introduce variable costs based on token consumption, model choice and usage patterns. As applications and agents scale, this can make expenditure harder to predict and allocate.
Unity Gateway provides usage visibility, cost attribution and budget controls to help organisations understand where AI spend is occurring. Its traffic management capabilities can also help teams manage consumption across supported model services.
Databricks also offers Smart Routing, which can select models and agent harnesses based on factors such as cost, performance and quality. Availability and supported configurations depend on the feature.
2. Strengthening AI security and governance
AI systems increasingly interact with sensitive information and external tools. Organisations need to control not only who can access a model, but also which tools an agent can call and what actions it is permitted to take.
Unity Gateway extends Unity Catalog governance into these runtime interactions. Organisations can apply permissions to governed AI assets and configure service policies to control requests and responses.
Depending on the configuration, controls can include filtering or masking sensitive information, restricting access to tools and recording activity for auditing.
This helps teams establish clearer boundaries around AI use without treating every model or application as a separate governance challenge.
3. Maintaining flexibility across AI models
The AI market continues to evolve, with new proprietary and open models offering different capabilities, costs and performance characteristics.
Unity Gateway supports access to Databricks-hosted models and external model providers, allowing teams to manage supported services through a common governance layer.
This gives developers more flexibility when selecting models for different tasks, while helping organisations maintain consistent oversight as their AI requirements change.
4. Improving visibility across AI activity
Understanding how AI is used is essential for managing performance, costs and risk. Without centralised monitoring, teams may struggle to identify which applications are consuming the most resources, investigate unexpected behaviour or understand how agents are interacting with tools.
Unity Gateway provides monitoring and usage information across its services, including request activity, token consumption, latency and cost-related data.
These insights can support operational decisions, help identify unusual usage and inform how AI services are managed over time.
How does Unity Gateway work with Unity Catalog?
Unity Catalog provides the governance foundation for data and AI assets within Databricks. It manages permissions and supports the discovery and governance of assets such as data, models, functions and tools.
Unity Gateway builds on this foundation by extending governance to the interactions that take place when AI services are used.
For example, Unity Catalog can control which users have permission to access a model or tool, while Unity Gateway applies configured policies to requests and manages traffic to the relevant services.
This distinction is important. Governing an AI asset is only part of the challenge; organisations also need to manage how that asset is accessed and used at runtime.
By bringing these capabilities together, Databricks provides a more consistent approach to governing data and AI across the platform.
What are the use cases for Unity Gateway?
Unity Gateway can support organisations looking to introduce or expand AI across different teams and workflows.
- Governing AI agents: Organisations can manage access to agents and the tools they use, helping establish consistent controls as agent-based applications become more widely deployed.
- Managing AI coding assistants: Engineering teams can use multiple coding agents and model providers. Unity Gateway helps centralise access controls, usage monitoring and cost management across supported coding workflows.
- Controlling access to enterprise tools: AI agents may need to interact with systems through Model Context Protocol servers and other tools. Unity Gateway can help restrict access to approved services and apply policies to these interactions.
- Managing multiple AI models: Where different tasks require different models, organisations can manage supported model services through a central layer and use available routing capabilities to balance cost, performance and quality.
- Monitoring AI expenditure: Teams can use usage and cost information to understand consumption across users, projects and applications, helping them make more informed decisions about AI investment.
Is Unity Gateway right for your organisation?
Unity Gateway is particularly relevant for organisations that are expanding their use of AI and need a more consistent way to manage access, security, costs and usage.
It may be worth exploring if your organisation:
- Uses multiple AI models or providers across different teams.
- Is introducing AI agents that interact with business systems and tools.
- Needs greater visibility of AI expenditure and consumption.
- Wants to apply consistent governance policies across AI services.
- Needs to monitor AI activity and maintain audit records.
- Wants to retain flexibility when choosing AI models and technologies.
Unity Gateway is a governance and management capability, rather than a complete AI strategy in itself. Organisations still need to establish clear use cases, suitable data foundations, security requirements and measures of success. Its value depends on how these capabilities fit into the wider data and AI environment.
How Telefónica Tech can help with Databricks and AI governance
Adopting AI at scale requires more than selecting the right models. Organisations need the data foundations, governance frameworks and engineering capabilities to put AI into practice responsibly.
Telefónica Tech helps organisations develop and implement data and AI strategies, build governed data platforms and identify opportunities to apply AI to business challenges.
For organisations exploring Databricks Unity Gateway, this means considering how AI governance fits into the wider architecture, including data access, security, operational requirements and cost management.
Whether you are establishing your AI foundations or expanding existing AI deployments, Telefónica Tech can help you assess your requirements and define a practical approach to implementation.
Want to explore how Databricks can support your AI strategy? Get in touch to discuss your data and AI requirements →
