How Trendyol scaled n8n to 1,000+ users and 700 active workflows in under a year

Background

Trendyol is the leading e-commerce platform in Türkiye, with offices in Istanbul, Amsterdam, Baku, Dubai, Riyadh, and Berlin, and a platform supporting 13+ languages and roughly 300,000 sellers. The company has been expanding into Central and Eastern Europe and the Gulf region, and employs over 10,000 people including blue-collar staff. The technology org alone is about 2,000 people across roughly 200 teams.

For the past year, Trendyol has been pushing AI and agentic workflows into every part of the business. A lot of that work now runs through n8n: a seller chatbot wired into the core product, a legal AI assistant in Slack, a search relevancy agent, AI code review, and a long tail of smaller automations built by individual teams. None of it existed twelve months ago.

The first deployment was a local proof of concept by Hüseyin Ulaş Head of Engineering. "I was looking for a tool that made creating workflows easy," he said. "When I saw the README, I was immediately impressed."

Challenge

Trendyol's technology org wasn't short on build power. Around 2,000 people across roughly 200 teams, plus an internal agentic platform of their own, and plenty of things teams wanted to ship. The problem was throughput. Anything new had to compete for engineering time and find space on a product roadmap. Even small things like a Slack message when a system event fires, or a daily summary of yesterday's numbers, mostly didn't make the cut. The non-technical side of the business had its own backlog and no real way to act on it. What Trendyol needed was something that let people, technical and non-technical, build their own answers to these problems without queuing for engineering time.

Solution

Hüseyin ran the POC locally, then deployed n8n on Trendyol's internal servers for a small pilot team. Within a month, the company had bought an Enterprise licence. "The community edition lacked logging and SSO, both of which were essential for our security requirements," Hüseyin explained. The first Enterprise licence was signed in June 2025.

After that, the rollout mostly ran itself. "n8n quickly became a massive hit within the company. I barely did any promotion. People saw it, they wanted to try it, and they instantly loved it," Hüseyin said.

To channel the demand, Merve Çolak Özkan and Barış Hasan Aras, another PM working on Trendyol's agentic transformation, ran an AI awareness programme with the HR learning team: around ten in-class sessions, and a community guild where builders swap solutions and help each other. "We always use real-world use cases we've already implemented here at Trendyol," Barış said. "The examples depend on the audience. If we're talking to the HR team, we focus on hiring processes; if we're with the technology team, we do a deep dive."

The platform is self-hosted, with around 200 projects, one for each technology team. Each project is scoped so users only see workflows and credentials inside their own project, with team managers as project admins. "We act as the main maintainer of n8n," Merve explained. "We implemented security regulations, including logging and using a vault for credentials. Each team is scoped to their own project, ensuring they can only create workflows and access credentials within their own area, keeping all other teams' information secure." The maintainer team meets daily with the security team to review updates and guardrails.

Building is decentralised on purpose. Anyone can build their own workflow, and for more complex projects the central workflow team takes ownership, but with a rule attached: the requesting team has to maintain the workflow afterwards. "We strongly encourage people to build and own their workflows. If they get stuck, they come to us for help,"

Three use cases worth highlighting

Three workflows stand out as the ones the team points to most often.

The seller chatbot is the biggest. It lives inside the existing seller panel, where sellers log in to upload products, set prices, and look at their analytics. Sellers can now ask it things like "what's my revenue over the last seven days?", and the chatbot routes the request through an orchestrator that classifies the intent and calls one of around ten n8n workflows behind it. SQL queries, and price analysis all run as actions, not just answers. The chatbot is currently live to roughly 10% of Trendyol's 300,000 sellers, with full rollout planned by the end of the year.

The legal AI bot is internal. It uses RAG over Trendyol's internal legal documentation and lives in Slack, where employees can tag the bot with a question before reaching out to the legal team directly. The international legal team validates answers and uses it to speed up their own responses. Behind the scenes, n8n retrieves answers from the knowledge base and compares responses across multiple LLMs.

The Search Relevancy Agent is now operational. Running periodically on low-CTR (Click-Through Rate) search terms across four storefronts—Romania, Greece, the UAE, and KSA—the agent evaluates the first 40 product results for each search query to generate a relevancy score. If this score falls below 80, the system automatically triggers a ticket for the respective team. This automation is designed to proactively detect search relevancy issues and streamline the resolution process, ultimately driving higher CTRs on the flagged terms. Looking ahead to the next phase, n8n will be integrated to execute corrective actions automatically, eliminating the need for human intervention.

Results

In under a year, n8n has gone from one developer's local POC to more than 1,000 active users at Trendyol, with 700 active workflows running in production and roughly half a million executions a quarter. Around 200 isolated projects exist, one per technology team, each with scoped access and credentials.

The qualitative impact is harder to measure but consistent. "If I have an idea, I would normally have to go to the product team, get it onto their roadmap, and wait for it to be prioritized," Hüseyin Ulaş, Head of Engineering, said. "With n8n, we can just try our idea and see the results immediately. It makes us more confident and faster: we don't have to wait to see if something works."

For the tech team, the savings are concrete. "In our technology department, we are constantly building small utility applications—things like 'if X happens in the system, send a Slack message'," Hüseyin said. "These are mundane tasks. They aren't hard to implement, but they consume time. With n8n, our technology teams can create these without any effort. The real benefit here is the time saved for our developers."

And for non-technical employees, the reaction tends to be the same. "When non-technical people see the drag-and-drop features and how easy it is to use, they are genuinely surprised," Barış said. "They start thinking about their daily tasks as flow diagrams. They can take their notes, drop nodes, and connect them. It's quite amazing for them to realize they can do these things so easily."