There is a category of research work that never makes it onto anyone's calendar. Check the overnight news on the watchlist. See if the margin story held up after the quarter. Reread the thesis when the price does something strange. None of it is hard. All of it depends on you remembering, at the right moment, on a day when you are busy with something else.
We have written before that the failure in investing is rarely intelligence. It is discipline. The checks are simple; doing them every time is not.
So we taught the analyst to show up without being asked.
A monitor is an analyst request on a schedule. You write it once, and it runs when you said it should.
Monitors are live in Investi today. Here is what shipped, and where this is going.
Write it once, pick when it runs
A monitor is the same request you would type into chat, saved with a calendar attached. "Summarize overnight news for my watchlist and flag anything that touches a thesis." Every weekday at 7am. "Has anything changed in the competitive picture for this company?" Every Friday. "Read the earnings release and compare it against what I wrote last quarter." Once, the morning after the report drops.
The schedule works the way a calendar event does. Run it once, every day, every weekday, every second Thursday, the last Friday of the month. You pick a start time in your timezone and the repeat rule follows from it. There is no cron syntax to learn and no rule language to get wrong.
The prompt is a full analyst request, so everything you can do in chat works here. Reference a company, a note, or a folder of notes with @, and the analyst reads them before it starts. A monitor that checks a thesis actually opens the thesis.
New monitorA monitor is a saved request with a calendar attachedEvery run is a real conversation
When a monitor fires, it starts a fresh chat. Not a notification, not a one line alert, a full analyst session with its reasoning and its data calls visible, exactly as if you had asked yourself.
That distinction matters more than it sounds. A run you can open is a run you can continue. If the Friday check on a company surfaces something odd, you are already in a conversation about it: ask the follow up, pull the filing, update the note. The monitor did the remembering and the first pass. The judgment part still starts where it always did, with you reading the work.
Runs collect under the monitor that produced them, and a small dot tells you when one has landed that you have not read yet. Catching up after a week away means opening a list, not reconstructing what you missed.
morning-briefEvery weekday · 7:00 AMTariff headline overnight touches your NVDA supply chain noteToday, 7:00 AMQuiet night, nothing crossed your thesesYesterday, 7:00 AMASML guidance cut, read against your semis noteTue, 7:00 AMEvery run is a chat you can open and continueThe analyst can set them up too
You do not have to build a monitor by hand. Mid conversation, you can just say it: "check this again every Monday and tell me if anything changed." The analyst drafts the monitor and shows it to you the way it shows every change to your research, as a proposal you approve before anything happens.
This is the same line we hold everywhere in the product. The analyst never quietly gains standing instructions any more than it quietly edits your notes. A schedule is a commitment to run in the future, which makes it exactly the kind of thing that should pass through you first.
Nothing runs on your behalf until you have seen it and said yes.
.monitors/nvda-weekly-checknew---starts_at: 2026-08-07T16:30:00-07:00rrule: FREQ=WEEKLY;BYDAY=FRtimezone: America/Los_Angeles--- Check whether anything changed in the competitivepicture for @NVDA and read it against my thesis.RejectApproveA monitor the analyst proposes still waits for your yesAnd a monitor stays an ordinary, editable thing afterward. Open it, change the wording, move it to a different day, pause it for earnings season, delete it when the thesis closes. It is your text on your calendar, not a black box.
The other half of monitoring
We should be honest about what this first release is and is not.
If you have used a broker app, you know the classic alert family: price crosses a level, volume spikes, a position moves four percent in a day, a 52 week high. Those are reactive triggers. They fire on market data, instantly, and they are useful precisely because they are dumb: a threshold either broke or it did not.
Monitors as they ship today are the other kind, proactive: the analyst goes out on your schedule and does the work whether or not anything moved. That order is deliberate. A price alert tells you that something happened. An analyst tells you whether it matters, against what you own, what you wrote, and what you said you were watching for.
But the reactive kind belongs in the same picture, and we are spending real time on how to get it there without bolting a generic alerts tab onto the side of the product. The interesting version is the two working together: the threshold as the trigger, the analyst as the response. Price crosses your level, and instead of a push notification you get a run that has already reread your thesis by the time you open it.
Your schedule firesThe analyst does the workA run, read against your notesPrice crosses your levelA bare push notificationThe same analyst run insteadProactive is live today; reactive triggers, same analyst, are nextThat part is not built yet. This part is.
The schedule was the missing piece. Discipline stops being a memory problem when the check runs whether you remembered or not.
Monitors are live now, next to skills and your notes. See how they work on the features page, or read Teach an AI analyst your investment process for the other half of making the analyst work the way you do.