---
title: "Notification Automation in 2 to 4 Weeks | JustAutomate"
description: "One rule instead of twenty alerts from every system. Thresholds, escalation, quiet hours and read receipts. Live 2 to 4 weeks after the workshop."
url: "https://justautomate.ai/en/services/notification-automation/"
lang: en
---

Process automation

# Notification automation with thresholds, escalation and quiet hours

Sending more alerts is not the point. The point is that every event has one owner, one threshold, one channel and a ladder to climb when nobody reacts. We agree all of that in the workshop, then put one rule in place of the twenty notifications each system sends on its own.

 Book a free consultation [See how the workshop works](https://justautomate.ai/en/services/process-workshop/)

- 72 hoursthe time the GDPR gives you to report a personal data breach to the supervisory authority, counted from becoming aware of the breach, not from the moment somebody opens the log. Article 33 of Regulation 2016/679
- 4 channelsSlack, Microsoft Teams, email and SMS. The channel follows the urgency of the event and the recipient, not the habit of whichever system sends the alert
- 2 to 4 weeksfrom the workshop to working notifications, tested on your own real events from the last few weeks

The problem

## Why does a team mute the notifications it asked for?

An invoice for EUR 40,000 sat 18 days past due, because everyone assumed somebody else was watching it. A lead that came in on Thursday afternoon waited until Monday for a reply. A manager heard about a technical fault at the moment the client asked for a project update. A campaign burned budget for a third day, because nobody looks at cost per acquisition over a weekend.

None of those companies was short of notifications. The CRM sends its own, the task system sends its own, the shop sends its own, monitoring sends its own, and they all land in the same inbox, in the same tone, addressed to everyone, with no threshold. A team learns to stop reading them inside two weeks. From that point an outage looks exactly like a newsletter in the inbox, which is worse than having no alerts at all, because it leaves the impression that something is watching.

Some deadlines come from outside the company and cannot be moved. [Article 33 of the GDPR (Regulation 2016/679)](https://eur-lex.europa.eu/eli/reg/2016/679/oj) gives you 72 hours to report a personal data breach to the supervisory authority, and it counts that time from the moment you become aware of the breach. If you find out about an outage from a client email on Monday, the statutory clock started on Friday evening, without you.

- An alert that matters is lost in a stream the team no longer reads
- A statutory deadline runs while nobody in the company knows the event happened
- Nobody knows whether anyone picked the alert up, because there is no read receipt
- People get messages at night about things that could have waited until morning

Why now

## What changed, so that notifications can finally be put in order?

The systems you already run can say out loud that something happened. A CRM, a task system, a shop and a payment gateway send a webhook the moment an event occurs, and BigQuery, SQL databases and Google Sheets can be queried on a schedule. So the decision about who needs to know and when moves out of the settings of a single tool into one place you control.

The second change is in how you pick a channel. Every channel has a different cost of attention: Slack or Microsoft Teams for things that happen during working hours, email for what can wait until morning, SMS for events that wake somebody up. The choice follows the recipient and the urgency, not whichever integration somebody switched on. One channel has its own page, because you ask about it most often: [automatic Slack notifications](https://justautomate.ai/en/automations/automatic-slack-notifications/) are one part of this, not the whole of it.

The third change is the AI agent. An alert says that something happened just now, but it will not answer the question of how many events like that you had this month and how many of them somebody picked up. BOND, our company AI agent, answers that from company data in 5 to 10 seconds, in a plain sentence, with no logging into five systems.

What changes

## What does the company get once the alerts are in order?

- ### An alert goes out only when somebody has to act
Every event passes through the same filter: who reacts, within what time and what action they take. An event with no owner or no action stays in the system for review on demand and sends nothing.
Three control questions in the workshop, instead of the default "switch everything on"
- ### An escalation ladder instead of hoping somebody notices
No reaction inside the agreed time lifts the case higher: from the person doing the work to their manager, from the manager to the board. A backlog stops depending on whether one person has their inbox open today.
Three levels of escalation, each with its own clock
- ### The channel follows the recipient and the urgency
An outage goes by SMS to whoever is on duty, an overdue invoice by email to the account owner, a team matter to Slack or to Microsoft Teams. The same mechanism, four different costs of attention.
SMS only for events on a list you approve in the workshop
- ### Quiet outside working hours
Alerts that are not urgent wait and go out as one summary at the start of the working day, following a calendar with your hours, your public holidays and your on-call rota. Urgent ones go out at once, at any hour.
One summary at 8:30 instead of twenty messages overnight
- ### A read receipt instead of guesswork
The recipient confirms with one click that they are taking the case, and the confirmation is stored with the time on it. No confirmation inside the agreed time triggers the next rung of the ladder on its own.
Who picked the event up and at what time, recorded on every alert
- ### One rule instead of twenty streams
We switch off the native notifications that duplicate each other, and in their place goes a condition that reads several systems at once. Instead of two separate emails you get one alert that already knows the case is serious.
A condition across several fields: invoice past due and no client contact for 14 days

Proof

## Do you watch events this way yourselves?

Yes. In our founders' agency the checks the team needs to know about are run by an AI agent, which reports the result itself. Below is the scale of that work, together with what the number does NOT prove.

Built for ourselves. These are not client deployments but implementations inside the founders' own companies. We say so plainly, because proof you cannot verify is not proof.

- ![JustIdea Agency](https://justautomate.ai/klienci/justidea.png)![JustBot](https://justautomate.ai/agenci/justbot-80.webp)JustBottheir agent
JustIdea Agency, a performance marketing agency of around 30 people
23
people from the company work with the agent
38
systems wired in permanently
112
processes run without a human
475
thousand lines of the agent’s code
The agent is wired into the CRM, the task system, the ad platforms, the monitoring of client websites and the data warehouse. It went live in March 2026 and 23 people use it today.
8 of 38 systems it works in

  - Google Ads
  - Meta Ads
  - HubSpot
  - ActiveCollab
  - Ahrefs
  - GA4
  - Search Console
  - BigQuery
Before
Whether something needed attention was decided by a person who went in and checked. Every check was a separate piece of work, and the result still had to be passed on to somebody.
After
The checks run in the background, and what reaches people is the result, not raw data to go through. A person steps in where the check found something.
29,854 of 64,903 operations in 30 days, or 46 percent, carried out with no human involved

And in your company

## What if our notifications already work?

Four sentences come up most often, and this is how we answer them.

- We have notifications in every system and they work.
The question is not whether they arrive, it is how many of them somebody opened this week. The first result of the workshop is subtraction: we usually switch off more existing notifications than we add new ones. What stays are the ones with an owner, a threshold and an action.
- We do not want to message people in the evening or at the weekend.
That is why the list of urgent events is short and you approve it yourself. Everything outside it waits and goes out as one summary at the start of the working day, following a calendar with your hours and your public holidays. We read the on-call rota and time off from the calendar, so an SMS does not reach somebody who is away.
- We do not use Slack, we live in Microsoft 365.
The mechanism is the same: an event from a source, a condition, a formatted message to a channel or to a person. We do this on Microsoft Teams, on email and on SMS exactly as we do on Slack. The channel is a decision about the recipient and the urgency, not about the tool we happen to like.
- Our events sit in five systems, one rule will not cover that.
In the workshop we set the source of truth for each event, so the same case does not fire two alerts from two systems. Where a field the condition needs is empty, or typed by hand three different ways, we say so plainly and we tidy up that one field first.

CRM

HubSpot, Pipedrive, Salesforce, Livespace

Delivery channels

Slack, Microsoft Teams, email, SMS, Telegram, webhook to any system

Team tools and tasks

Google Workspace, Microsoft 365, Notion, ClickUp, Jira, Asana

Sales, shop and finance

Shopify, WooCommerce, PrestaShop, BaseLinker, Xero, sevDesk, Comarch Optima

Data and metrics

BigQuery, Google Sheets, Google Analytics, SQL databases

Automation layer

Make, n8n, Zapier, custom API integrations

## Channel, threshold and escalation for typical events

This is what a single row of a notification policy looks like: the kind of event, how it is delivered and what happens next when nobody reacts. The timings are examples, because you set them yourself in the workshop.

| Kind of event                                               | Channel and mode                                       | What happens with no reaction                                               |
| ----------------------------------------------------------- | ------------------------------------------------------ | --------------------------------------------------------------------------- |
| System or integration failure                               | SMS to whoever is on duty, at once, including at night | after 15 minutes with no confirmation, SMS to the second person on the list |
| Suspected personal data breach                              | SMS and email to the person responsible, at once       | after an hour it goes to the board, because the GDPR deadline is 72 hours   |
| Invoice 14 days past due                                    | email to the account owner, once a day in a summary    | after 30 days the case goes to the finance channel and to the manager       |
| Lead with no first contact                                  | Slack or Teams to the owner, during working hours      | after 2 hours the sales manager is told                                     |
| Task past its deadline                                      | message to the person doing it, on the project channel | after 2 days to the project manager, with a list of what is overdue         |
| A metric outside its band, cost per acquisition for example | summary to the board channel, once a day               | two days in a row outside the band lifts it to whoever owns the budget      |
| A request waiting for approval                              | email in the daily summary                             | after 24 hours a reminder, copied to the manager                            |

The rows are a starting point. A rule with no agreed owner and no agreed action does not go into the rollout, because it will be muted within a month anyway.

How we work

## Five steps from noise to a notification policy

Each step ends with something you can judge before you decide on the next one.

1. ### Workshop and inventory
We write down the events somebody at your company has to know about, and separately we count the notifications already going out: how many a day, from which systems and how many of them anyone actually opens. Without that second list, new alerts only add to the noise.
2. ### The notification policy on paper
For each event we set the owner, the channel, the threshold, the escalation ladder, the way receipt is confirmed and the quiet hours. We also agree what data may go into the body of an alert and what stays behind a link to the record. What comes out is one document that stands on its own without us.
3. ### Rollout in parallel
The rules start alongside your current notifications, on a test channel with a few people. We run them against real events from the last few weeks, so you see the actual number of messages a day rather than the number from the plan.
4. ### Switch over and turn off the duplicates
Once that checks out we move the alerts to their proper channels and switch off the old notifications that duplicate them. We set up monitoring of the rules themselves, because a silent failure of alerts is more dangerous than having none: the team assumes that if nothing arrived, nothing happened. Somebody on your side learns to change a threshold and an owner without us.
5. ### Measurement at 30 and 90 days
We look at two numbers: how many alerts went out and how many of them somebody confirmed. A rule with zero reaction gets a higher threshold or goes, because an alert nobody reads spoils all the others.

Before and after

## How the team's week changes

### Before

- Alerts from five systems land in one inbox and have an "archive" rule on them
- The backlog comes up for the first time at the weekly review
- Nobody knows whether anyone picked a case up, so the question goes round the channel
- People get messages at night about things that could have waited until morning

### After

- The channel and the inbox carry only events with a threshold, so the team reads them
- No reaction lifts the case higher on its own, inside the agreed time
- Every alert shows who picked it up and at what time
- Outside working hours one summary goes out, and urgent things go at once

The first rules are visible on the day of the switch. We judge the thresholds, the channels and the escalation ladder at 30 days against what actually went out, and again at 90 days.

For the person signing this off

## What the operations director gets out of it

You stop being the channel through which news of problems travels. An event reaches the person responsible on its own, and when they do not react inside the agreed time, the case moves up on its own. That takes away the role of the person who has to remember everything and chase everybody.

The second thing carries more weight in a board conversation. You know how many events of a given kind there were in a month, how many of them were confirmed and how many had to be escalated, because the alerts and the confirmations are recorded. In an audit, or after a personal data breach, you show a trail of decisions rather than an account from memory.

- The notification policy is written down, so it does not depend on one person in the company
- Statutory deadlines watched by a rule, for example the 72 hours in Article 33 of the GDPR
- Rules run in parallel, so the old notifications keep working until testing ends
- Reaction measured at 30 and 90 days, not judged from a corridor conversation

What it costs

## The workshop has a fixed price, the rollout is quoted after it

The process workshop costs **EUR 3,500 net per week** and runs from one to seven weeks depending on scope. For notifications one week is usually enough, because the scope is narrow: the list of events, the thresholds, the channels and the escalations. The rollout is quoted after the workshop, once we know how many source systems are in play and whether the data the conditions need is complete.

The cost of running the rules goes straight to the vendor, Make, Zapier or an SMS gateway for example, or you pay nothing at all when n8n runs on your own infrastructure. Where an AI agent is in scope, the model provider charges EUR 0.005 to 0.07 per query. No commission on top from us.

- ### Process workshop
EUR 3,500 net
per week, usually one week for a notification engagement
A self-contained piece of work. It ends with a notification policy you can execute without us, or hand to anybody else.

  - A list of events, each with an owner, a threshold and a channel
  - The escalation ladder, the rules for confirming receipt and the quiet hours
  - A list of notifications to switch off, because they duplicate or nobody reads them
  - A tooling recommendation and the metrics to measure at 30 days
 Start with the workshop
- Most popular
### Workshop and notification rollout
Quoted after the workshop
scope follows from the number of events, channels and source systems
The option most companies choose. Working notifications 2 to 4 weeks after the workshop, tested on your own real events.

  - Everything in the workshop
  - Rules in n8n, Make or through the API, wired into the CRM, tasks, the shop and the warehouse
  - Delivery to Slack, Microsoft Teams, email and SMS, split into urgent and summaries
  - Escalation when nobody reacts, read receipts, quiet mode outside working hours
  - Monitoring of the rules themselves and training for the person who will change thresholds
 Talk about a rollout
- ### Notifications and an AI agent
Quoted after the workshop
for companies where somebody asks about the state of things every day
Everything in the rollout plus BOND, which answers questions about events and reactions in a plain sentence.

  - Everything in the notification rollout
  - AI agent wired into the CRM, the task system and the data warehouse
  - An answer from company data in 5 to 10 seconds, with no logging into systems
  - Daily and weekly summaries written by the agent, not by a person
 Ask about the agent

If the workshop shows that switching off a few notifications and setting two thresholds in the tools you already run is enough, we say so and show you how. The workshop ends with a document that stands on its own without us, and the engagement does not move on to a rollout you do not need.

## Frequently asked questions

What is notification automation?

It is agreeing and building a policy: which events deserve an alert, who gets it, at what threshold, in which channel and what happens when nobody reacts. The rules pick events up from your systems and deliver them to Slack, Microsoft Teams, email or SMS. The result is fewer messages than before the rollout, but ones that somebody reads.

How is this different from the notifications we already have in every system?

Native notifications send what the vendor of the tool anticipated, usually with no threshold, no choice of recipient and no knowledge of your other systems. One rule on our side sees several sources at once, so it can join an overdue invoice to the absence of any client contact. If native notifications are enough in your case, we say so in the workshop.

How do you decide which event deserves an alert?

We ask three questions: who has to react, within what time and what action they take. An event with no answer to any one of them stays in the system and is available on demand, but sends no notification. That is the most common reason the workshop produces a shorter list of alerts than the one it started with.

How does escalation work when nobody reacts?

Every rule has a time to react and a next recipient. No confirmation of receipt within that time moves the case one rung up: from the person doing the work to their manager, then to the board. You set the levels and the timings in the workshop, because for an outage that can be 15 minutes and for an overdue invoice 30 days.

Can notifications be switched off outside working hours?

Yes, and it is a separate decision for each rule. Events on the urgent list go out at once, at any hour, and the rest wait and go out as one summary at the start of the working day. The rule knows your working hours, the public holiday calendar and the on-call rota, so an SMS does not reach somebody who has time off.

Which channel do you pick: Slack, Teams, email or SMS?

Urgency and recipient decide. We keep SMS for events that wake somebody up, Slack or Microsoft Teams for team matters during working hours, and email for what can wait until morning and needs a trail in the inbox. One event can use two channels, when the first one gets no confirmation.

How long does a notification system take to build?

One week of workshop and 2 to 4 weeks of rollout, with tests on real events from the last few weeks. The first working rule usually goes in within 14 days of the workshop. With more source systems, or with data that has to be tidied up first, it stretches to 6 weeks.

What happens when the notifications themselves stop going out?

Monitoring of the rules is standard: a failed run reports itself to the technical channel and to whoever is on duty, and the rule catches up on the events from the outage window once it is back. A silent failure of alerts is more dangerous than having none, because the team assumes that if nothing arrived, nothing happened.

## Let us check what your team finds out about too late

The free consultation takes 30 minutes. You tell us which systems you run and which event you last heard about from a client rather than from your own tooling. We show you which rules can be set up straight away, which notifications are worth switching off first and where the data for a condition is missing. You do not have to share access to any system at this stage.

We run at most two rollouts at a time, so that someone is actually sitting on each of them.

[Book a free consultation](https://justautomate.ai/en/contact/)

**PS.** If somebody at your company has set up an inbox rule that archives alerts from the CRM, that is exactly this problem. A notification with no threshold and no owner turns into noise, and then you hear about an overdue invoice or an outage from a client. An event with a threshold, an escalation and a read receipt cuts that down to minutes. Writing those events down is where the workshop starts.

## Related

- [Automatic Slack notifications](https://justautomate.ai/en/automations/automatic-slack-notifications/)
- [Reporting automation](https://justautomate.ai/en/services/reporting-automation/)
- [CRM automation](https://justautomate.ai/en/services/crm-automation/)
- [n8n automation](https://justautomate.ai/en/services/n8n-automation/)
- [BOND, business AI agent](https://justautomate.ai/en/services/bond-business-ai-agent/)
- [Process workshop](https://justautomate.ai/en/services/process-workshop/)

---

**JustAutomate** · contact: contact@justautomate.ai · form: https://justautomate.ai/en/contact/