Zendesk CSAT Trend Reporter

What it does

Zendesk CSAT Trend Reporter is a customer-success intelligence agent. It pulls CSAT satisfaction ratings from Zendesk (last 30 days by default), buckets them by week, computes a moving average, groups negative comments into actionable themes (slow response, unresolved issues, poor communication, etc.), ranks support agents by score, and conditionally fires a Slack alert when overall CSAT drops below a configurable threshold. Every analysis step is deterministic; the LLM is used only for narrative synthesis, with a template fallback when no key is set. Built for customer-success and support-ops leaders who want early warning on quality decline instead of manually pulling Zendesk reports.

Example prompts

• Run the full CSAT trend analysis and alert Slack if it's declining
• Show me the agent leaderboard by CSAT score
• What are the top negative feedback themes this month?
• Check if we need to trigger a manager alert right now

Tools (8)

Tools the agent exposes — your AI client calls them automatically when it needs them.

• run_full — Run the full CSAT analysis pipeline: collect Zendesk ratings, compute weekly trends, extract negative themes, optionally send a Slack alert.
• get_csat_trends — Given raw CSAT data, compute weekly trend analysis with moving averages and return the full domain output.
• get_agent_leaderboard — Given raw CSAT data, return per-agent CSAT scores sorted by score with threshold flags.
• get_negative_themes — Extract and rank the top negative feedback themes from unsatisfied CSAT comments using keyword analysis.
• check_alert_condition — Check if current CSAT data meets the condition for triggering a manager alert.
• plan_inputs — Plan the inputs for a tool: returns questions, schema and a ready-to-edit example.
• discover_intent — Map a natural-language query to the best matching tool and return its plan_inputs.
• list_capabilities — List the agent's available tools, credential slots, and supported intents.