Reports | Yaplet

Where to find it

Team → Reports in the left menu. This menu group used to be called Analytics; the page and its address are unchanged.

The page is one screen with three pill tabs across the top. Each tab is its own view, so only the charts you are looking at are loaded.

Web analytics

Visitor behavior, traffic sources, engagement metrics, custom events, and an activity heatmap.

Live chat

Agent statistics, customer ratings, geographic insights, and agent chat activity over time.

AI & usage

AI credit spending, event usage, and a per-model breakdown of credits and costs.

Each tab has its own date range, so you can look at a different period in each one without disturbing the others.

A privacy-friendly overview of your website traffic — sessions, pageviews, bounce rate, referrers, devices, geography, custom events, and an activity heatmap. This tab also shows a live count of online users currently on your site.

Live chat

Your support team's performance: per-agent response times and star ratings, the countries your visitors connect from, and a timeline of when agents join chats or the system requests help. The activity chart can be filtered by widget and switched to counting each chat once.

Opening a single agent from this tab shows their own page. If your organization has Agent monitoring, that page also carries an Agent monitoring (last 30 days) card — how many of that agent's answers contradicted your knowledge base, how many are still unreviewed, and across how many checked replies.

AI & usage

What your AI features cost. An AI usage spending chart, an event usage engagement chart, and an AI usage by model table breaking down credits and cost across LLM, embedding, reranking, TTS and voice.

How far back the numbers go

Web analytics reaches back about four months. The tab is built from daily summaries that Yaplet keeps rather than holding raw browsing history forever, and the raw page-view and session records behind them are removed on a nightly schedule once they pass that age. In practice this means the date picker on that tab refuses a start date older than 119 days, and shifts the range forward with a short message if you try.

That window does not apply to the other two tabs — but Live chat is not unlimited either, because the conversations behind the figures are deleted in their own right: three years after a conversation's last activity while the organization has an active subscription, and 120 days after it when it does not. Live chat reaches back as far as the conversations you still have, and no further. The full schedule is on GDPR & Data Privacy. The AI & usage tab has its own picker allowing any period up to 93 days.

The AI & usage tab no longer has a Fetch type control. That "fast/slow fetching" selector, which used to reach older historical data, was removed — the tab now has a date range picker and nothing else.

Shared controls

Grouping

Where a chart supports it — the Web analytics tab, and the agent activity chart on the Live chat tab — a period selector next to the date picker controls how the data is bucketed: hourly, daily, weekly or monthly. It only offers the groupings that make sense for the range you picked: a single day can only be hourly, a week only daily, a month daily or weekly, a quarter weekly or monthly, and anything longer is monthly.

Date ranges

Date ranges are set per tab, and one tab can carry more than one picker. Web analytics has a single picker, up to 365 days long, with no start date older than 119 days. AI & usage has a single picker allowing up to 93 days. Live chat has three, set independently: one on the agent statistics panel, one on the top countries panel (90 days at most), and one above the agent activity chart — the first two open on the last month, the third opens on today.

Tips

Watch agent performance. Check response times and ratings on the Live chat tab regularly — that is where a training need or a top performer shows up first.

Keep an eye on cost. The AI & usage tab is where credit spending becomes visible before it becomes a bill.

Know where your audience is. The geography breakdowns tell you which markets are actually growing, which is worth knowing before you plan support hours around them.

Automate with AI: You can also pull analytics data using Copilot or any MCP-compatible client. Try "Show me agent performance for the last 30 days" or "Compare this month's chatbot analytics with last month". See the Analytics Tools reference for all available operations.