Case study

Business process automation

We moved the core processes of a premium-brand appliance online store onto automation: AI agents run SEO and content, n8n handles orders and customer communication, and round-the-clock monitoring sends alerts to Telegram. The store has no in-house developer. We don’t disclose the company’s name: the project is under NDA.

About the project

Client

Premium-brand appliance online store (name under NDA)

Industry

E-commerce: home appliances and hair care devices

Services

Business process automation, SEO, technical site support

Period

October 2026 — present

The task

The task and starting point

The store runs on WordPress and WooCommerce. There is no dedicated developer and no content team: tech, marketing, order handling and site monitoring all depended on manual work.

In autumn 2026 the site went through a security incident, and organic traffic dropped. The store needed to grow SEO, keep every order and learn about any site problem right away — without hiring staff for each task.

What we did

Which processes we automated

Every process is built as a chain: an event on the site or a schedule → processing by an agent or an n8n workflow → a result on the site, in a spreadsheet or as a Telegram alert. Changes to the live site are approved by the owner.

Marketing and SEO: content run by AI agents

AI agents run promotion through a “strategy → brief → copy → fact check → SEO review” chain. Agents don’t invent specs: only manufacturer documentation and store data, with a source for every number.

  • A keyword set of 21,709 queries and 86 blog clusters collected automatically via the Keys.so API
  • The blog launched on November 6, 2026; articles are prepared in batches and published on schedule — 4 per week
  • SEO copy and FAQ with structured data for categories, product card copy, and a manuals section based on official PDFs
  • 12 new products added to the site in one day with official images and specs
  • New and updated pages are submitted to search engines automatically via IndexNow
  • Every two weeks — a rankings and clicks report (Keys.so, Google Search Console) in a shared spreadsheet

Sales: orders through n8n

Every new WooCommerce order goes straight to an n8n workflow via webhook. The manager doesn’t have to watch the admin panel or copy data by hand.

  • The order lands in a Google Sheets order log the moment it is placed
  • A new order alert arrives in Telegram
  • Order data is available to n8n workflows for further processing

Customer communication

Messaging with customers is part of the automation too and uses the same order data: the manager works from a ready order record instead of collecting details across different systems.

  • Customer and order details are pulled in automatically
  • Requests don’t get lost between the site, the spreadsheet and the messenger

Security: monitoring with alerts

After the incident we set up automatic site checks. Alerts go to a Telegram bot, the same alert repeats at most once every 6 hours, and an “alerts cleared” message follows the fix.

  • Every 15 minutes: integrity of key files, PHP files in uploads, new PHP and .htaccess files, admin account changes, firewall events
  • Every 15 minutes: what Googlebot and YandexBot see and whether the home page responds
  • Daily: core and plugin checksums, SSL certificate expiry and a digest
  • External uptime monitoring: 7 checks every 5 minutes

Results

What changed

  • Marketing, sales, customer communication and security run without an in-house developer or content team
  • The blog publishes 4 articles a week on schedule; the catalog grew by 12 models in one day
  • Every order reaches the spreadsheet and Telegram without manual copying
  • The owner learns about a site problem within 15 minutes: we tested the alert with a test file — both the alert and the “cleared” message arrived

Stack

n8n, AI agents on Claude Code, WordPress and WooCommerce, Google Sheets, Telegram bots, UptimeRobot, Keys.so API, Yandex Metrica, Google Search Console, IndexNow.

Methodology

The list of automations and dates comes from the project’s working repository and work log as of 2026-11-12. Content and monitoring figures come from configuration and change history. We don’t publish revenue or order data.

We don’t disclose the company’s name or site address under the NDA.

Sources

  1. The project’s working repository: work log, monitoring and workflow configuration, checked 2026-11-12. Not published: the client is under NDA.
  2. Confirmation from the project owner on the sales and customer communication automation, 2026-11-12.

Next step

Let’s talk about your processes

Tell us which tasks in your business are still done by hand — we’ll show what to hand over to agents and n8n first.