How we saved a client 50 hours a month

The client runs an online store with sports accessories — roughly 400 active SKUs, 3,000 orders a month and a team of four people. Customer support, order processing and administrative processes were taking up a disproportionate amount of their time. They came to us with a simple question: “Where are we losing the most time and how do we automate it?”

Step 1: Analysis (weeks 1–2)

Before we deploy anything, we spend time understanding how the company actually works. Not how it thinks it works — but how it really works every day.

With this client we tracked incoming e-mails for two weeks. The results were clear:

• 38% of e-mails concerned the status of a shipment or an order
• 22% of e-mails were requests for a return or a complaint
• 18% of e-mails were questions about the availability and parameters of products
• 12% of e-mails were billing matters
• 10% of e-mails were genuinely complex cases requiring human judgement

In other words: 90% of e-mails were predictable and repeatable. Ideal ground for automation.

Alongside e-mails we mapped the internal processes: how orders are processed, how communication with warehouses and carriers works, how daily reports are created. We found that the team spent an average of 3 hours a day copying data between systems — purely manual work with no added value.

Step 2: Designing the solution

Based on the analysis we designed three modules that address the biggest pain points:

Module 1: E-mail assistant. It processes incoming e-mails, classifies their intent and automatically answers questions about shipment status (connected to the carrier), product availability (connected to the warehouse) and the returns procedure (automatic template + ticket creation). More complex cases it hands over to a specific team member together with the context.

Module 2: Order processing automation. Every new order automatically triggers a series of actions: warehouse notification, stock check, carrier assignment, CRM update, entry into the reporting table. Without a single click.

Module 3: Daily reporting. Every morning at 7:30 the owners receive an automatic overview: the previous day's orders, stock levels for the top 20 products, open complaints and the average response time of customer support.

Step 3: Deployment (weeks 3–5)

The deployment was gradual, without interrupting operations. We started with the e-mail assistant in “shadow mode” — the assistant suggested the replies, but the team approved them. After a week the accuracy of the replies was above 94%, so we switched to fully automatic mode.

The order processing automation was deployed in the second week — connecting Shoptet, the warehouse system and the carrier took two days of setup and one day of testing.

Reporting was the simplest part — two scenarios in Make.com, one day of setup.

Results after 90 days

Here are the numbers the client shared with us after three months of operation:

• 52 hours a month saved on customer support and administration
• The average response time dropped from 5.5 hours to 4 minutes
• 94.3% of e-mails are handled by the assistant without the human team stepping in
• Customer satisfaction (measured by simple star feedback after a query is resolved) rose by 18%
• The team shifted from a reactive e-mail queue to proactive activities — improving the range on offer, communicating with suppliers, preparing campaigns

What we learned

This project confirmed several things for us that we now apply with every new client:

Analysis before design. Without a detailed view of what the team actually does, you would automate the wrong things. Start with data, not intuition.

Shadow mode saves your nerves. Launching automation at full scale right away is an unnecessary risk. A week in shadow mode uncovers errors before they reach customers.

Results show up quickly. This client saw the first measurable results within two weeks of deploying the e-mail assistant. The full results were apparent within 60 days.