MTS Web Services Switched from RAG to AI Agent for Initial Support Ticket Triage

MTS Web Services built an AI agent to automatically process technical support tickets — and abandoned its RAG-based search system, which required engineers to formulate queries correctly.

Why they built RAG — and why it wasn't enough

The SRE team of the second-line support for MTS Web Services' advertising technology cluster works daily with four data sources: Jira with incident history, Confluence with documentation, system logs, and databases. When analyzing complex incidents, engineers spent considerable time searching for similar cases and analyzing past solutions.

The first solution was a RAG system: data from Jira and Confluence were combined into a single searchable knowledge base supporting natural language queries. Engineers could ask questions — and the system returned relevant documents and incident records. The effect was notable, and the team received positive feedback.

However, over time a structural limitation became apparent. RAG is a search tool: for it to work, the user must know what to search for. When analyzing incidents, this condition doesn't hold: the cause of the problem is unknown, it's unclear what data is needed and in what order to request it. As a result, the tool remained underutilized precisely where help was needed most — at the start of the investigation.

The second problem was the inability to measure real impact: positive feedback existed, but objective metrics — how much time was saved per ticket, how MTTR changed — couldn't be obtained.

How the agent takes on the first stage

"If the system can already search for information and analyze what it finds, why not have it handle the first stage of incident investigation entirely?" —

Vladimir Drobot, SRE Lead at MTS Web Services.

Instead of a passive tool waiting for a query, the team built an agent that conducts investigations autonomously. The agent receives a ticket as input and independently performs the first stage:

• Analyzes the incident description and formulates an investigation plan
• Independently determines what data is needed for analysis
• Formulates and executes search queries to Jira and Confluence
• Aggregates information from multiple sources
• Provides the engineer with structured preliminary analysis including hypotheses about causes

The difference from RAG is fundamental: the agent doesn't wait for the right question from a human — it decides what to search for based on ticket context.

What this means

The MTS Web Services experience demonstrates a key distinction between a helper tool and an autonomous agent in a corporate environment. RAG supplements an engineer at the stage when they already know where to look. The agent takes on the work that consumed the most time — initial hypothesis formation and data collection — and frees up the costly attention of L2 specialists for real analysis. This shows that agentic architecture is already practical today for internal corporate tools, not just public products.

Why are companies implementing AI for technical support?

To automate initial ticket processing and save engineer time. MTS switched to an AI agent because the previous RAG system required specialists to know how to formulate search queries correctly.

How does MTS's AI agent analyze tickets?

The agent processes four data sources: incident history in Jira, documentation in Confluence, system logs, and databases of the advertising technology cluster.

What data sources does the AI agent use in technical support?

MTS's AI agent works with four sources: Jira with incident history, Confluence with documentation, system logs, and databases.