Cockpit AI briefings for managers

You run a shared inbox — support, sales, or info@ — and several people answer from it. You don't have the hours to open every thread and judge whether the team is on top of things. The MailDesk Cockpit gives you a single screen that already knows the answer, and an AI Insights panel that turns that screen into a short, readable briefing you can act on in under a minute. Instead of skimming conversations, you read one paragraph and a short to-do list — and you know exactly whe

Available in: Cockpit integration (maildesk_mail_client_cockpit). The Cockpit module builds on MailDesk Pro and uses the same AI provider you configure for Pro. Without Cockpit there are no manager briefings; without a configured Pro AI provider the written briefings stay disabled.

You run a shared inbox — support, sales, or info@ — and several people answer from it. You don't have the hours to open every thread and judge whether the team is on top of things. The MailDesk Cockpit gives you a single screen that already knows the answer, and an AI Insights panel that turns that screen into a short, readable briefing you can act on in under a minute. Instead of skimming conversations, you read one paragraph and a short to-do list — and you know exactly where to look first.

The dashboard, in one glance

The Cockpit is an operational dashboard for the people who own a shared mailbox. Across the top it shows live KPI cards — how much came in, how much needs a reply, how much is overdue, and your average first response time — pulled straight from your real MailDesk accounts. Every number is clickable: open the exact email, the thread, or the backlog row right there, without leaving the screen.

That's useful on its own. The AI Insights panel beside it is what turns those numbers into a decision.

What the AI Insights panel does

The panel sits in the Cockpit's side area. A small badge shows your access tier (for example manager), and below it you get a focused set of actions — each one answers a question a busy manager actually asks.

• Generate daily briefing — a one-page written summary of your queue: the state of the queue, the risks and anomalies worth knowing about, and 2–4 recommended actions for today. The text writes itself in live, and you can copy it straight into your stand-up notes or a message to the team.
• Customer escalation risk — ranks the customers most likely to escalate, so you can reach out before the complaint arrives. The ranking is based on open threads, overdue threads, urgency wording, and how long the oldest message has been waiting.
• Rebalance workload — proposes how to spread open work more evenly across the team when one person is drowning and another has room.
• Anomalies (last 30 days) — flags any metric that is unusually high or low compared with its recent normal, each with a severity badge, so a quiet spike never goes unnoticed.

Why managers love it

A manager watching several mailboxes does not have time to read every thread. The daily briefing condenses the whole day into a paragraph and a short action list. The escalation-risk and anomaly tools point you straight at the customers and metrics that need attention first — no opening, no skimming, no guesswork. And because the statistical tools never call an AI model, they are free to run and always available, even on day one before a provider is set up. You spend your morning acting, not investigating.

Run a daily briefing

This is the heart of the page — three clicks from open mailbox to a briefing you can paste into your stand-up.

1. Open MailDesk → Cockpit from the MailDesk menu.
2. In the side panel, find the AI Insights block. The badge in its corner shows your tier (for example manager) — that's how you know which actions you'll see.
3. Click Generate daily briefing.
4. A briefing panel slides in and the text streams in live, headed by a quiet Thinking… indicator. While it is still writing you can press Stop. When it's finished, the header reads Ready and you can Copy the text or Close the panel.

What the briefing looks like

The briefing arrives in three short sections — State of the queue, Risks & anomalies, and Recommended actions today — built entirely from your real KPIs and any detected anomalies. It's concrete: a couple of sentences of context, the risks worth knowing, and two to four things to do today. Read it, copy it, share it, move on.

The free tools — no provider needed

These three run entirely on the Cockpit's own data and never contact an AI provider, so they work immediately, cost nothing, and are great for a quick check before you've even set AI up:

• Anomalies (last 30 days) compares the latest day against a 30-day baseline and lists the metrics that are unusually high or low, each with a severity badge. (Before a provider is configured, this same button reads "Detect anomalies (statistical, no LLM)".)
• Customer escalation risk lists the customers most at risk, each with a score and a severity.
• Rebalance workload proposes how to spread open work across the team.

Who can see what — the four tiers

Cockpit uses a four-tier access model, and each tier includes everything the tier below it can do. The written briefings (and the rest of the manager toolkit) are reserved for Manager and above — that keeps the serious, account-wide picture in the right hands.

What you need first

• The Cockpit module installed. It builds on MailDesk Basic, MailDesk Pro, and the Automation module (plus Odoo's Resource module) — Odoo installs all of these for you automatically.

• A working AI provider — but only for the written briefings. Cockpit reuses the very same provider you already configured for MailDesk Pro, so there's one key for both. See Set up your AI provider. The statistical tools (anomalies, escalation risk, rebalance, mailbox health) need no provider at all.

• At least one mailbox account visible to the manager, with some message history so the KPIs have something to report.
• MailDesk works on Odoo 17, 18, or 19 — the steps above are the same on each.

Daily usage limits (quotas)

Every AI action is counted per user, per day, per operation — a fair-use cap that keeps costs predictable. When someone reaches the limit for an operation, MailDesk simply asks them to try again tomorrow or have an administrator raise the cap. Administrators have no daily limit.

When something looks off

For provider-side errors (an invalid key, quota exhausted at the provider, an unreachable local server), see the AI troubleshooting page.

Related

• What MailDesk AI can and cannot access
• AI troubleshooting
• Realtime & synchronization architecture