Intelligence
Automated fault analysis across your charger fleet: grouped issues with severity and auto-resolve tracking, the detected-faults event list, AI insight cards with trends, fault alert recipients, and the email delivery log.
The Intelligence section watches your network for problems so you find them before drivers report them. It combines two detection sources into one picture: AI analysis of failed charging sessions and diagnostic errors reported by the chargers themselves (offline, voltage problems, and similar). The section has three areas: Charger health, Notification settings, and Email delivery, plus a Zone monitoring shortcut that opens the external monitoring tool where configured.
Charger health: the overview
The top of the page gives you the selected period at a glance, with a period picker (day, week, month):
- Top insights: AI summary cards describing the most important things happening in your fleet, with an "Updated" timestamp.
- Health index and stats: a large health index plus Analyzed sessions, Successful sessions, Fault rate, Faults, and Users affected. The health index and fault rate show a trend arrow compared to the previous period, so you can see whether things are improving.
- Graphs: faults by reason, faults by origin, and top lists for Failures by location and Failures by charger.
Issues: one row per problem, not per event
The Issues tab groups repeated faults on the same charger (and connector, when known) into a single issue, so one flaky charger reads as one problem with a count, not thirty scattered rows. Both detection sources feed the same issue: AI-analyzed session faults and charger-reported diagnostic errors.
Each issue shows:
- Severity: Minor, Warning, or Critical. Charger hardware and firmware faults start at Warning, everything else at Minor, and an issue climbs one step for each of: 5 or more users affected, 3 or more occurrences within 72 hours, or 3 or more faults in a row without a successful charge. Hover the severity pill to see exactly why an issue has its level.
- Issue: what is wrong, on which charger, at which location.
- Origin: where the fault points (charger hardware, vehicle, user, and so on).
- Trend (14d): a 14-day sparkline of occurrences.
- Events and Users: how often it happened and how many distinct users were affected.
- Last seen.
Filter with the Active, 30 days, and All chips, by CPO, location, origin, or charger model, or search. When everything is healthy, the Active view says so: "No active issues. All chargers are healthy."
Auto-resolve
An active issue resolves itself after 5 successful sessions on the affected connector since the last fault (the pill shows the progress, for example 3/5). A new fault reactivates the issue. Resolved issues show an Auto-resolved pill with a frozen severity.
The issue detail page
Click an issue to open its full picture:
- A stat row: Occurrences, Users affected, First seen, Last seen, and Successful sessions since last fault with the auto-resolve target.
- Port timeline (last 14 days): the connector's actual status history, so you can see whether sessions run between faults and how long the port stays unavailable.
- Occurrences per day (last 14 days): a histogram of how often it hits.
- Latest AI analysis: the root cause, a recommendation, an expandable Learn more with the full analysis, and the key events of the latest occurrence, linking to the source session. Issues reported only by the charger itself (no session to analyse) say so plainly.
- All occurrences: every event with date, session link, and user.
Detected faults: the raw event list
The Detected faults tab keeps the flat, per-session view: date, fault type, session, charger, location, recommendation, and origin. Diagnostic events from the chargers appear here too. Use it when you want the individual events rather than the grouped picture.
Notification settings
Configure who receives fault alert emails and for which locations. Add recipients with Add recipient, choosing an admin and either all locations or a subset. Each row can be edited or deleted.
Email delivery
The Email delivery page is a log of the transactional emails the platform sent on your behalf (confirmation codes, invoices, fleet and home charger invites), covering the last 5 days:
- Columns: Date, Recipient, Type, Subject, Status.
- Statuses include Delivered, Opened, Clicked, Permanent fail, Temporary fail, Spam complaint, and Unsubscribed.
- Expand a row for a plain-language explanation of what the status means and the raw server response; delivered rows include a spam-folder hint for "the customer says they got nothing" cases.
- Filter by email type and status, or search by recipient email.
Access to this page is permission-gated; ask your account owner if you need it.
Common questions
Why does one charger show thirty faults? Switch to the Issues tab: repeated faults on the same charger are grouped into one issue with an occurrence count and a trend, which is the intended way to read them.
What do the severity levels mean? Minor, Warning, Critical. Hardware and firmware faults start at Warning; an issue escalates with 5+ users affected, 3+ occurrences in 72 hours, or 3+ faults in a row without a successful charge. Hover the pill for the exact reasons.
When does an issue disappear from Active? After 5 successful sessions on the affected connector since the last fault, it auto-resolves. The 30 days and All views keep showing resolved issues.
How do I get emailed about faults? Intelligence → Notification settings → Add recipient, picking the admin and locations.
A customer says they never received their confirmation code or invoice. How do I check? Intelligence → Email delivery: search their email address. The status shows whether it was delivered, opened, or failed, and delivered-but-unseen usually means the spam folder.
Where did Diagnostics go? Charger-reported diagnostic errors now feed the same Issues and Detected faults views as the AI session analysis, so there is one place to watch instead of two.
Updated 5 days ago