Product pack

The "See the proof" link opens the list that formed the number, with the filter already applied. The action button goes straight to the place to resolve it. The table exports to CSV via the page button.

The finding carries the subject's screen. Whoever lacks access to that screen does not receive the alert — neither in the bell nor by e-mail. And whoever cannot See Values receives the finding with the number masked, not a meaningless finding.

Each finding has a signature (the type plus the identity of the target). While it is open, the watch does not record another identical one — not even when two servers run the same cycle in the same minute.

The Silence button applies to that type of finding, for you or for the whole company. It is on purpose: whoever only has the option of turning everything off ends up losing the alert that mattered along with the one that was a nuisance.

The finding only reaches those with access to the subject's screen. A tool that does not declare the screen — nor explicitly declares that the subject belongs to everyone in the company — does not generate findings: the silence is deliberate, because sending without knowing to whom has already sent a financial alert to the entire company. And the quantity in the alert is what the tool declares it found; "nothing found" does not become an alert, even when the query responds with a row saying so.

Notices marked "(operador)" are not about your company: they are the platform watching itself. Every 15 minutes the Tenka Operator reads repeated errors on a screen, screens above the agreed time, screens that had data and opened empty, delayed background routines, stuck e-mail or WhatsApp queues, searches with no result, Tenkai's "I couldn't find", abandoned forms, repeated clicks and help that did not help. Each notice brings the screen, how many times, how many people and accounts (number only, never the name) and the time window.

The notice gets a severity by effect: S1 blocks money or a first-use screen and the on-call is warned at once; S2 is a wrong function with a workaround or a slow screen and goes into the daily report; S3 is polish. S1 and S2 open a ticket in the house tenant with "how to reproduce" and "where the trail is" — the "Open the ticket" action leads to it. The operator never fixes, restarts or touches data: it only opens and warns. A daily report sums up what broke, what got slow and what nobody found.

When someone in the account tries to open a screen and is stopped by the group or the plan, administrators get an alert with the person’s name, the screen and the reason, plus a button to see their access. One alert per person, screen and day. The error messages people see on screen are recorded too, and Tenkai answers what broke today.

When the same problem continues the next day, today's notice is linked to the ticket that is already open ("already open") instead of opening another; a new ticket is only created after the previous one is closed. The daily notice carries only S1 and S2; polish (S3) goes, listed, into the weekly report, which comes out once a week with what broke, what got slow and what users could not find. With an AI key, Tenkai writes a summary paragraph above the facts; without one, the report comes out with the facts only and says so. The operator now also warns when the same screen was denied to several people for the same reason and when someone's AI usage was far above normal, always with numbers, never with names.

From reading tools that already existed and warned no one: posting anomaly, accounts due this week, contract about to expire, indicator off target, budget overrun and minimum treasury balance. The watch runs at the end of the day and only records what is new.