Geckoboard

Claude can investigate why CSAT fell. It shouldn’t have to calculate CSAT from thousands of Zendesk tickets first. But that’s what often happens when an AI assistant only has access to raw records. This week, we launched the Geckoboard Metrics MCP, a connector that gives AI assistants like Claude and ChatGPT access to accurate, real-time, pre-calculated Zendesk metrics. You can ask: “Why did CSAT fall last week?” “Which queue is driving the increase in first response time?” “How does this quarter compare with the previous one?” Geckoboard calculates the metrics. Your AI assistant investigates what changed, combines the numbers with context about your business and creates the report, briefing or resource you asked for. It can also bring in metrics from platforms like Aircall, HubSpot, Pipedrive and Shopify, giving it a broader view of the customer journey. It’s like having a support data analyst inside your AI assistant. The Geckoboard Metrics MCP is now live and is included on every Geckoboard plan. See it in action. Link in comments.

First reply time rose, but that doesn’t necessarily mean the team got slower. In this example, an AI assistant uses the Geckoboard Metrics MCP to investigate eight weeks of support data. It finds that part of the increase came from a shift toward email, a slower support channel. After accounting for that change, the remaining delay is concentrated in early-morning Tier 1 email. Breaking the metric down turns a broad concern about team speed into a focused investigation of one channel, tier and time of day.

The leadership meeting starts in 20 minutes! What does the support update need to cover? In this example, an AI assistant uses the Geckoboard Metrics MCP to compare demand, SLA performance and backlog health. Demand and SLA performance have recovered, while 34 Tier 1 web-form tickets remain at least seven days old. The result is a concise weekly brief that captures the recovery and makes the unfinished work clear.

Many teams still spend hours every week building reports to answer questions that take seconds to ask. “Why did win rate fall last month?” Answering it usually means opening a reporting tool, building a new report or query, trying different breakdowns and comparing periods until you find an explanation. An AI assistant should make that easier. But it can only do the job properly if it has access to calculated metrics. Connect one directly to an API and it may receive thousands of individual records instead. It then has to retrieve all the relevant data, navigate pagination and API limits, and decide how to calculate the metric itself. That can be slow, expensive and error-prone. The Geckoboard Metrics MCP takes a different route. Geckoboard calculates the metrics before passing them to your AI assistant. You can ask: “Investigate the fall in win rate last month and show me where it was concentrated.” Your AI assistant can start with the metric, investigate what changed and turn the findings into a report or another useful resource. You get the answer without building another report, and your AI assistant doesn’t have to reconstruct the metric from raw records. See how the Geckoboard Metrics MCP works. Link in comments.

AI assistants can make incredible data analysts. The problem is, most of them don’t have access to the metrics they need. Connect an AI assistant directly to a business app and it may receive thousands of raw records rather than a straightforward answer. It then has to find the relevant data, navigate pagination and API limits, and decide how to calculate the metric itself. That can be slow, expensive and error-prone. Today, we’re launching a different approach. The Geckoboard Metrics MCP is a connector that gives AI assistants like Claude and ChatGPT access to accurate, real-time, pre-calculated business metrics. Ask: “Why did win rate fall last month?” “What caused support volume to increase last week?” “Summarize last week’s performance and flag anything unusual.” Geckoboard handles the metrics. Your AI assistant investigates what they mean, combines them with context about your business and turns the findings into something useful. It can create reports, dashboards, presentations and custom tools. It can also analyze metrics from different platforms in the same conversation. It’s like having a data analyst inside your AI assistant. The Geckoboard Metrics MCP launches today with more than 15 integrations and is included on every Geckoboard plan. See how it works. Link in comments.

Getting ahead of customer frustration changes how a support team prioritises everything. At Spotler UK, Tom Mercer (Head of Support) has brought real-time sentiment data powered by Stylo directly in front of agents using Geckoboard TV dashboards. It puts the team on the front foot. They can spot risk early and act before issues escalate. Instead of reacting to angry tickets after the fact, they’re able to re-prioritise work, de-escalate sooner, and make calmer decisions based on how customers are actually feeling in real time. We’ve now partnered with Stylo to make this setup available to other Zendesk teams. If you’re already using Geckoboard, you can bring Stylo’s AI-driven sentiment signals straight into your dashboards. And with our partner offer, you get 3 months of Stylo free to use alongside Geckoboard. Link with full details in the comments 👇