Automations — Complete Guide

Automations put your recurring research on autopilot. Instead of opening the platform to re-run the same report, refresh the same grid, or ask the same question every morning, you set it up once and let it run — on a fixed schedule, or the moment a company you follow has an earnings call or other corporate event. When a run finishes, the results are emailed to whoever you choose (and kept in the dashboard), so the work is waiting for you instead of on your to-do list. Find it under Research Workspace → Automations.

​

What it is

An automation is a saved instruction that says “run this task, on this trigger, and deliver the result here.” Each one has three parts:

• A task — the actual work: generate a Report from one of your templates, refresh a Grid, or run an AI Chat prompt against the data sources you choose.
• A trigger — when it runs: on a recurring schedule (e.g. every weekday at 9:00), or in response to a corporate event (an earnings call or investor-relations event for the companies you watch).
• A delivery — where the result goes: emailed to one or more recipients as a finished file, and always saved in the dashboard so you can review run history any time.

Automations are how you turn one-off research into a standing process: a morning briefing, a post-earnings analysis for every name on your watchlist, a weekly sector grid, a Monday data digest.

​

When to use it

Reach for Automations when:

• You run the same task on a rhythm — daily, weekly, monthly, hourly — and want it done without you.
• You want to react to earnings or IR events across a list of companies without watching the calendar yourself.
• You want results delivered to a team by email on a schedule, in a finished format.
• You’ve already built a Report template or a Grid and want it kept fresh automatically.

Use something else when:

• You need a one-time answer right now → run it directly in Reports, Grids, Agent Studio, or Chat. (Build it there first, then automate it here.)
• You’re still designing the report template or grid → create and refine it in Reports or Grids; an automation just runs what already exists.
• You want an open-ended, interactive research session → use Agent Studio or Chat.

​

How to use it

Creating an automation is a short, guided sequence. You can leave most settings at their defaults and just fill in what matters to you. Once created, every automation has its own detail view where you can:

• Turn it on or off with the Active toggle (an inactive automation never runs).
• Run it now with the Run button — a one-off run that happens immediately, separate from the schedule. (For an event automation that uses the dynamic event company, you’ll be asked which watched company to use for the manual run.)
• Edit any setting, or delete it.
• See its run history — every run with its result (success or failed), timing, and a short message — which refreshes on its own while you watch.

​

Capabilities & key choices

​

The three task types

​

The two triggers

Schedule — a recurring timer. Choose a frequency and time:

• Weekdays (Mon–Fri), Daily, Weekly (pick the day), Custom days (pick any combination of weekdays, e.g. Mon/Wed/Fri), Monthly (pick the day of the month), or Hourly (every 1, 2, 3, 4, 6, 8, or 12 hours).
• Set the time of day, and the timezone the schedule should follow — type to search for a city or zone (your runs always honor that timezone, so daylight-saving shifts don’t move them). The builder shows a plain-English summary of what you picked — e.g. “Every weekday at 09:00” — which also appears on the review and detail screens.

Event — run in response to a corporate event for companies you watch:

• Event type: Earnings Events (a company’s earnings call) or Investor Relations Events (non-earnings IR events such as presentations, investor meetings, and shareholder meetings).
• Companies to watch: add up to 200 companies by ticker (type to search, or bulk-import a list). The automation fires when any watched company has a matching event.
• Timing relative to the event: run a set time before or after the event — presets range from 3 days before to 3 days after (including 1 hour before, 20 minutes before, at the event, 15 minutes after, and more), or set a custom offset in minutes, hours, or days.
• When an event fires for one of your watched companies, the task runs for that specific company automatically — so a single event automation watching a whole watchlist produces one tailored run per company, per event.

​

AI Chat: response depth

An AI Chat automation runs at one of four depths — match it to how much thinking the task needs:

• Instant — the fastest; immediate answers for simple lookups and quick checks.
• Balanced — fast, everyday research and checks; good for standard questions and summaries.
• Agent (default) — full multi-step analysis with tools; the right choice for most scheduled work, including in-depth write-ups.
• Deep Research — the most thorough; multi-source investigation for weekly deep-dives and peer comparisons. It takes the longest.

​

AI Chat: apply skills

You can have an AI Chat automation apply one or more skills on every run — the reusable playbooks that tell the assistant how you like a task done. Switch between AllMind skills (the built-in library) and your own saved skills, search either list, and select as many as you need. The skills you pick shape how each scheduled run approaches your prompt, and the number you chose is shown on the review and detail screens.

​

Attach your own files (AI Chat and Grid)

Beyond the connected sources, you can upload your own documents directly in an AI Chat or Grid automation — a model, a memo, a spreadsheet, a PDF. They’re saved with the automation and read on every run (for a Grid, they’re analyzed together with the refreshed grid when its prompt runs). Files upload with a progress indicator, and you can’t create the automation until the uploads finish. Most common formats are accepted (PDF, Word, Excel, PowerPoint, CSV, text, and images), up to about 12.5 MB each. (Report automations don’t take file attachments — they run from your saved template.)

​

Delivery format

When email delivery is on, you choose what’s attached:

• Report → PDF (default), Word, or both.
• AI Chat → Word (default), PDF, or both.
• Grid → always an Excel file; add an AI Grid Prompt (see below) to also receive a PDF summary.

​

Grid summaries (the AI Grid Prompt)

For a Grid automation, you can add an AI Grid Prompt — an optional prompt the assistant runs over the refreshed grid (and any files you attached) once it finishes, delivered as a short PDF alongside the Excel file. Start from a ready-made template (Executive Summary, Risk Flags, Earnings Preview, Peer Comparison, Trend Analysis, or Action Items) or write your own. It turns a raw data refresh into a narrative you can read at a glance.

​

What it can access & produce

Reports and Grids draw on whatever their underlying template or grid already uses — the platform’s full market, fundamentals, filings, transcripts, and research data, exactly as when you run them by hand. For an event-triggered report, the triggering company is supplied automatically as the report’s subject. AI Chat can be pointed at a rich set of sources — turn on only the ones you want it to use: A few sources need a one-time connection in your settings first (Google Drive, OneDrive, SharePoint, Gmail, Outlook); Snowflake uses your organization’s authorized tables. If a connected source has been disconnected, the automation asks you to reconnect it before saving. For Gmail and Outlook you can restrict which email domains the automation may read. When the list is long, you can search and filter your sources (all, included, connected, or needs-action) and search within Data Room folders and Snowflake tables. Beyond these sources, an AI Chat automation can also apply your skills and read files you upload to it — both are set up in the AI Chat step (see Capabilities & key choices above). What automations produce:

• Report → a finished PDF and/or Word report, emailed and saved so you can open it later.
• Grid → your Grid refreshed in place (so the saved grid itself stays current) plus an Excel export, and an optional PDF summary when you add an AI Grid Prompt.
• AI Chat → a written answer as a Word or PDF document, emailed and also kept in your chat history.

Every run’s outcome is recorded in the automation’s run history, and you get an in-app notification when it completes.

​

Tips & best practices

• Build the piece first. Perfect your report template in Reports or your grid in Grids, then automate it — an automation runs what already works.
• Name automations for their job, not their mechanics (“Monday energy digest” beats “Grid 3”).
• Use one event automation for a whole watchlist. Add all the tickers you care about (up to 200) to a single earnings automation rather than creating one per company.
• Mind the timing offset. “15 minutes after” an earnings call gives the event time to happen; “1 day before” is useful for previews. Choose deliberately.
• Leave a report’s period blank to always analyze the latest data on each run.
• Match AI Chat depth to the stakes — Agent for most scheduled work, Deep Research for the heavy weekly deep-dives, Balanced for light checks, and Instant for the quickest lookups.
• Add an AI Grid Prompt to Grid automations so recipients get a readable summary, not just a spreadsheet.
• Attach the files a run should always read — a model, a mandate, or a methodology doc — so every scheduled run works from the same materials.
• Use Run now to test. Trigger a one-off run to confirm the output looks right before relying on the schedule.
• Turn automations off when you don’t need them rather than deleting — you keep the setup and its history.

​

Limits & things to know

• An event automation can watch up to 200 companies.
• Schedules can be as frequent as you like — down to hourly — within the frequencies the builder offers; all runs follow the timezone you set.
• Event look-ahead is about 90 days. The system finds the next matching event within roughly the coming quarter; upcoming triggers appear shortly after you create or save an event automation, and are rechecked periodically. If no matching event is found in that window, the automation shows “no upcoming triggers” until one appears (you can still use Run for a one-off).
• Runs are asynchronous and can take a while — from under a minute to around half an hour for the deepest work. You don’t need to keep anything open; you’ll be notified and the result will be waiting.
• Failed runs aren’t retried automatically. A failure is recorded in run history with a short reason, and the automation simply runs again at its next scheduled time. You can press Run to retry immediately.
• Inactive automations never run, and turning one off (or deleting it) stops all future runs.
• Run history shows recent runs in the detail view and refreshes on its own.
• Email delivery is optional — choose Dashboard Only and the results are still produced and saved, just not emailed.
• You can attach your own files to AI Chat and Grid automations — most common formats, up to about 12.5 MB each. They’re read on every run, and an automation can’t be created until its uploads finish.

​

How it works with other features

Automations sit on top of the rest of the platform and keep its outputs flowing on a schedule:

• Reports — build a report template once, then have an automation generate it on a cadence or around earnings. Automations are how report templates become recurring deliverables.
• Grids — point an automation at a saved grid to keep it refreshed with current data, and email the result (with an optional AI summary).
• Agent Studio / Chat — an AI Chat automation is essentially a scheduled agent run: the same kind of prompt, depth, and tools, but on a timer or an event.
• Data Room & Document Search / Library — an AI Chat automation can read selected Data Room folders and saved research documents, so scheduled analysis is grounded in your own materials.
• Data Sources & integrations — connected sources (Snowflake, Google Drive, OneDrive, SharePoint, Gmail, Outlook, Carbon Arc) become inputs an AI Chat automation can use on every run.
• Live Investor Relations / events — the corporate-events calendar is what Event triggers watch for earnings and IR events.

In short: other features create the work; Automations decide when it runs and where the result goes.

​

Example workflows

​

1. Automatic post-earnings analysis for your whole watchlist

Goal: get a tailored analysis of every company you follow, minutes after it reports.

1. In Reports, build (or pick) a report template for post-earnings analysis.
2. In Automations, create a Report automation and select that template.
3. Choose an Event trigger → Earnings Events, and add your watchlist (up to 200 tickers).
4. Set the timing to 15 minutes after the event.
5. Turn on Email, add your team, and choose PDF.

Now, whenever any watched company reports, everyone gets the analysis for that specific company, automatically — no calendar-watching required.

​

2. Weekly sector grid with a written summary

Goal: a fresh sector comparison and a readable takeaway every Monday morning.

1. In Grids, build a grid across your sector’s companies and the metrics you track.
2. In Automations, create a Grid automation and select that grid.
3. Choose a Schedule trigger → Weekly, Monday, 8:00, in your timezone.
4. Turn on Email and add an AI Grid Prompt — e.g. the Executive Summary template.

Every Monday the grid refreshes itself and your inbox gets the updated Excel plus a one-page PDF summary of what changed.

​

3. Monday data digest from your own systems

Goal: a weekly briefing that blends your internal data with market context.

1. Connect Snowflake (and set up any Data Room folders you want included).
2. In Automations, create an AI Chat automation with a prompt like “Pull the latest metrics for our top 10 accounts, summarize the week-over-week changes, and flag any account whose usage dropped more than 20%.”
3. Set the depth to Agent, and turn on Snowflake (with the right tables) plus Financial Data and any Data Room folder.
4. Choose a Schedule trigger → Weekly, and deliver as Word.

Each week you get a written digest grounded in your own numbers — assembled while you were away.

​

Common questions

​

Getting help

For help setting up an automation or designing a recurring workflow, reach AllMind support through the in-app support option or your account team.