> ## Documentation Index
> Fetch the complete documentation index at: https://docs.usequotable.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Routines

> Save instructions once and Quotable Agent carries them out for a brand on its own, on a schedule or when visibility, mention rate or share of voice moves.

> [!NOTE]
>
> **Plans:** All plans, [limits](https://docs.usequotable.com/workspace/plans-and-limits.md#how-credits-work) vary · **In the app:** Sidebar → Agent → Agents

A **routine** is a set of instructions Quotable Agent carries out for a brand without you in the chat. It reads your tracking data, researches the web, uses the skills you attach and writes a report. It runs on a schedule, when a metric moves, or when you click **Run now**.

In the sidebar, open **Agent**, then **Agents**: the app lists routines as agents. Routines belong to the brand you have selected.

## Create a routine

Click **New routine**, or open the arrow next to it for the other ways:

> [tab] Describe it to Quotable Agent
>
> Say what should happen and when, for example "every Monday at 9, summarise last week's visibility and the competitors that gained". Quotable Agent drafts the routine and shows it as a card with **Create**, **Edit first** and **Discard**. Nothing is saved until you confirm.

> [tab] Draft it in the side panel
>
> The same draft, in the Quotable Agent panel next to the page you are on.

> [tab] Write it yourself
>
> Opens an empty routine form.

> [tab] Start from a template
>
> Pick a template below the list. It opens the form filled in, and nothing is created until you save. See [Templates](#templates).

You can also ask for a routine in any Quotable Agent chat. It drafts it with the same card.

## What a routine holds

| Field | What it does |
| - | - |
| **Name** and icon | Up to 80 characters. |
| **Instructions** | What to do, in your own words, up to 8,000 characters. Be specific about the period, the metrics and the length you want. |
| **Skills** | Up to three [skills](https://docs.usequotable.com/actions/skills.md), applied to every run. |
| **When it runs** | A schedule, metric triggers, or both. See below. |
| **What it may change** | **Read only**, **Asks before changes** or **Changes without asking**. See [What a run may change](#what-a-run-may-change). |
| **Depth** | **Standard**, **Thorough** (about twice the credits) or **Deepest** (several times the credits). |
| **May stop to ask you** | When on, a run that needs a decision only you can make waits for your answer. |
| **Email me each report** | The report also arrives by email. It is always in the run's history. |

## When it runs

Turn on one or both switches under **When it runs**.

### On a schedule

At a set time in the workspace's time zone:

| Cadence | Example |
| - | - |
| Daily | Every day at 09:00 |
| Weekdays | Every weekday at 09:00 |
| Weekly | Every Monday at 09:00 |
| Monthly | Monthly on the 1st at 09:00 |

A monthly day from the 29th to the 31st runs on the last day of shorter months. The form shows the next run before you save.

### When a metric moves

A trigger watches one metric for the brand: **Visibility score**, **Mention rate** or **Share of voice**. Choose a condition:

* **drops by** or **rises by** an amount, in points or as a **% of the average**
* **falls below** or **rises above** a value

Compare against **the previous day**, or the **3-day**, **7-day** or **14-day average**. A routine can have up to three triggers, and any one of them starts a run.

Triggers are checked after each complete day of tracking, and fire at most once that day. Partial days are ignored. **falls below** and **rises above** fire on the day the line is crossed, not every day after. A new routine starts watching from the next complete day, not from data that came before it.

A routine with triggers and no schedule shows **Waits for a metric** in the list.

## What a run may change

Every run can read the brand's tracking data and the web. What it may change depends on the setting:

| Setting | What happens |
| - | - |
| **Read only** | Reads your data and writes a report. Nothing in Quotable or a connected app changes. |
| **Asks before changes** | Stops at each change and waits for you to approve it in the run. |
| **Changes without asking** | Makes changes through your [connected apps](https://docs.usequotable.com/integrations/apps.md) on its own. Use it only for instructions you trust. |

Every template starts as **Read only**.

## Runs

Open a routine to see its runs. Each run is a Quotable Agent conversation: the instructions, then the answer, with the steps it took shown as it works. You can reply in the run to ask a follow-up.

A run has one of these statuses:

| Status | Meaning |
| - | - |
| **Queued** | Waiting to start. |
| **Running** | Working now. |
| **Needs you** | Stopped for your answer or approval. You get a notification. |
| **Done** | Finished. |
| **Failed** | Stopped with an error. Run it again. |
| **Skipped** | Did not start, for example because generation is paused after a failed payment. |

Each run shows how it started: **On schedule**, **Run now** or **Metric change**. Run history is kept, and is not removed with the 7-day purge of ordinary chats.

A routine runs one at a time. **Run now** is refused while a run is still going: "A run of this routine is already under way."

## Manage routines

* **Pause** or resume a routine with the switch on its page. A paused routine keeps its settings and history.
* From **More**: **Edit everything**, **Change with Quotable Agent**, **Duplicate** and **Delete**.
* The list has **Mine** and **Paused** filters and a search box.

A routine runs as the person who created it. Members with access to the brand can see its routines; only its creator can open its runs and edit it. If the creator leaves the workspace or loses edit access to the brand, its runs are skipped. **Duplicate** it to run it as yourself.

Creating, editing, pausing and running routines needs edit access to the brand. See [Team and roles](https://docs.usequotable.com/workspace/team-and-roles.md).

## Templates

| Template | Category | When | What it does |
| - | - | - | - |
| **Weekly visibility brief** | Reporting | Mondays 09:00 | Last 7 days against the 7 before: visibility, mention rate, share of voice, prompts and sources that moved, three actions. |
| **Competitor watch** | Competitors | Thursdays 09:00 | Competitors ranked by change in share of voice, the prompts and pages carrying the top two, and untracked brands appearing in answers. |
| **Source gap sweep** | Sources | Tuesdays 09:00 | The ten domains cited for competitors but not for you over 14 days, and the three most realistic targets. |
| **Sentiment and facts check** | Reputation | Fridays 09:00 | Negative or mixed answers about the brand, and claims that look wrong or outdated, with their sources. |
| **Monthly opportunities review** | Reporting | 1st of the month, 09:00 | What was marked done on the [Opportunities](https://docs.usequotable.com/actions/opportunities.md) board and whether it moved anything, what is stuck, and what to do next. |
| **Visibility drop investigation** | Alerts | When visibility drops by 5 points against the 7-day average | Finds the engines, prompts, competitors and cited pages behind the drop. |

Change any of it before you save.

## Credits

Every run uses [credits](https://docs.usequotable.com/workspace/plans-and-limits.md#how-credits-work) from the workspace's monthly allowance. **Thorough** and **Deepest** use more. A run that the workspace cannot cover is refused before it starts. When generation is paused after a failed payment, scheduled and triggered runs are skipped. See [If a payment fails](https://docs.usequotable.com/workspace/plans-and-limits.md#if-a-payment-fails).

## Next steps

> [card] [Quotable Agent](https://docs.usequotable.com/actions/quotable-agent.md)
>
> The assistant that carries out every run.

> [card] [Skills](https://docs.usequotable.com/actions/skills.md)
>
> Reusable instructions to attach to a routine.

> [card] [Connected apps](https://docs.usequotable.com/integrations/apps.md)
>
> Let a routine work in an app your brand connected.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.