Alerts
People do not watch dashboards. Alerts post what changed to Slack, Microsoft Teams or email, and only when there is something to say. A week with nothing to report stays quiet. Every configured channel gets the same message.
Set up
Section titled “Set up”Slack: add an Incoming Webhook to the channel that should receive alerts (Slack, Apps, Incoming Webhooks, Add to channel), paste the URL under Settings, Alerts & reports and press Send a test.
Microsoft Teams: in the channel, create a Workflow of the kind “post to a channel when a webhook request is received” (or a classic incoming-webhook connector where those still exist) and paste its URL. The alert arrives as an Adaptive Card.
Email: list the addresses. The alert is a short email with the same lines and a link. It comes from Vercatus’s own address, so the recipients may want to allow-list it.
The URLs are stored with the workspace and used for nothing else. Admins can change them; analysts and viewers can see that alerts are configured but not the URLs.
What triggers an alert
Section titled “What triggers an alert”Each kind has its own switch under Settings, Alerts & reports.
| Event | Default | Why |
|---|---|---|
| The site audit found new critical or high findings, or findings that got worse | on | The reason to interrupt someone. |
| AI crawlers newly refused (or newly served); pages that stop returning 200 | on | Usually a firewall rule or a deploy. The sooner known, the cheaper. |
| A run fails or pauses | on | Someone has to resume it. |
| Fixes confirmed by the last run | off | Good news; turn it on if the channel likes it. |
| Every completed run, even with nothing to report | off | A heartbeat. |
| A question tracked word for word changed | on | An engine’s answers to a question you track word for word started or stopped naming or citing the site, two answers in a row. Webhooks receive prompt.changed. |
| An engine changed what it cites or rewards | on | A confirmed engine change that touches your sites. Webhooks receive engine.drift. |
| A competitor overtook a site | on | A competitor’s page score from the audit passed the site’s. Webhooks receive competitor.changed. |
| A search crawler slowed down on the pages that earn | on | Fetched less often before any loss shows in the answers. |
| An answer engine quotes what a site no longer says | on | A stale answer to correct. |
| A play’s verdict landed | on | A Won or Lost ground read (once for each state; a read before day 45 is marked provisional) and the final day-45 verdict whatever it says. It says how far answers moved, with the range and the grade. Webhooks receive verdict.read. |
| Answer Share moved clearly | on | Once a day at most, read from the continuous poll: your Answer Share, a topic’s or a competitor’s over the last 7 days against the 7 before, only when the move is beyond its range (the same rule the app uses for a clear change). A week with no clear move sends nothing. Webhooks receive visibility.changed. |
| A required disclosure went missing | on | A disclosure you defined was on a page at the last check and is gone now. |
| An agent task that passed now fails | on | The weekly trial of a buying task failed where it passed the week before. |
Each alert names the site, lists up to five items, and links to where they are. Audit findings are compared by their stable identity, so a finding already known does not alert again; only what is new, worse or gone does.
The access watch
Section titled “The access watch”The daily access watch uses the access-changes switch: a crawler newly
refused or served on a watched page, robots.txt or llms.txt changed, or a watched page gone, the day it happens.
Outgoing webhooks receive it as access.changed.
Crawler traffic
Section titled “Crawler traffic”With crawler traffic connected, the daily access watch also reads the crawlers’ own
requests: an AI crawler refused on half or more of yesterday’s requests after a quiet week (or the reverse, served
again), and a crawler that used to come and stopped, said once until it returns. Both go out under the access-changes
switch and as the traffic.changed webhook.
AI-referred traffic
Section titled “AI-referred traffic”With Google Analytics connected, each site is checked once a week: the sessions AI
assistants sent in the last complete week against the average of the four weeks before. A drop of more than 30% goes
out under the severe-findings switch, with the two figures and the week; weeks with very few sessions are left alone,
since a small number moves for no reason. Outgoing webhooks receive it as traffic.changed. The figure is a floor, as
everywhere it appears: assistants that strip the referrer arrive as direct, and Google’s AI answers as organic search.
Report delivery by email
Section titled “Report delivery by email”Under Settings, Alerts & reports, Report delivery, list the people who should receive the run report after every completed full run, with the run report and the board brief attached as PDF when the boxes are ticked. The monthly brief is set up on the same page.
Elsewhere
Section titled “Elsewhere”Plays can become issues in Linear or Jira, your own systems can subscribe to events over signed webhooks, and your CMS can tell Vercatus when something shipped: see Linear and Jira, Webhooks and CMS and deploy webhooks.