Since Enhance is built on an API-first architecture, integrating these capabilities is seamless, as nearly every action can be triggered via API.
While we don’t offer an MCP server to our customers at the moment, we are currently rolling out similar features in our customer portal. Our existing AI chat agent (inside our customer portal) is integrated with our Knowledge Base to handle operational queries and has real-time access to customer account data, including services, plan utilization, payments, and support tickets. The agent autonomously queries the necessary data based on user input using tool calls that communicate with our internal API, which wraps Enhance’s APIs and other services.
Furthermore, we are developing a feature that enables the agent to analyze PHP error logs for targeted troubleshooting. This is already partially functional in our staging environment and includes the ability to perform actions such as restarting PHP containers - always subject to user approval.
That said, we are cautious about exposing a direct MCP interface for customer-side agents. Providing external agents with direct action capabilities opens the door to misconfigurations and unintended consequences that could increase our support overhead. By keeping the AI agent within our own controlled environment, we can ensure it is finely tuned and operates within strict guardrails - level of governance that is much harder to maintain with third-party agents connecting via MCP.
