From the field · Event notes

How does Yargıcı use AI? Notes from the conversation at the Türkiye AI Summit

·6 min read·Skyloop Cloud

In the online programme on day two of the Türkiye AI Summit 2026, Tanıl İnci, Project Manager at Yargıcı, answered questions from Beyza Mensur, Business Development Manager at ZZeti. The subject was Yargıcı AI, the project Yargıcı has built on ZZeti Zeka.

Below we have gathered what was said under a few headings. The recording of the conversation, which is in Turkish, is included in the article.

The starting point: reaching data without waiting for a report

As Tanıl İnci described it, Yargıcı treats AI not just as a tool that fetches data but as part of its digital transformation and its daily decision processes. The first focus was to answer management's questions quickly and to reach information held in different systems faster.

The aim was not to keep data confined to Excel files and classic dashboards: managers should be able to reach data and analysis by asking in natural language, without waiting for someone to prepare a report. According to İnci, Yargıcı AI has largely delivered that. What matters is not only a fast answer, but reaching the same data with the right business rules and in a common language.

Why ZZeti Zeka and Skyloop?

Yargıcı spoke to several firms along the way. İnci said one of the advantages that stood out in going with Skyloop was its command of the Nebim ecosystem and its work with Nebim. For Yargıcı the ERP and the data warehouse are critical; the basic expectation was correct and fast integration with the existing infrastructure.

At the start of the project an analysis layer was built with ZZeti Zeka for Nebim V3 data. Alongside the Nebim database and data warehouse, company-specific data structures and business rules were brought into scope. Sales, product, store and inventory data became queryable on the same platform.

The collaboration went beyond standard reports: product photos were made accessible at platform level and added to product-based reports, and work was done on reaching and interpreting document content. In İnci's words, what mattered was a solution partner able to work together around the company's needs, in addition to a ready product.

Yargıcı AI project — in conversation with ZZeti
Customer Story
Yargıcı

Yargıcı AI project — in conversation with ZZeti

Tanıl İnci
Tanıl İnci
Yargıcı — Project Manager
Beyza Mensur
Beyza Mensur
ZZeti — Business Development Manager

Türkiye AI Summit (TRAI), Day 2 • Nov 2026

From one use case to the whole company

The first use case was querying sales data in natural language. The scope grew over time:

  • Sales, product and campaign analysis
  • Inventory and store performance analysis
  • Anomaly detection: making an unexpected change in an indicator visible
  • WhatsApp and Teams integration: access from the channels employees already use; usage over WhatsApp in particular rose markedly
  • A second solution, also running on local infrastructure, that supports employees with questions in their daily work
  • Agent automations: daily target-achievement rates are sent to the relevant people by WhatsApp; every Tuesday the weekly sales performance for the management meeting is prepared by AI and e-mailed to managers

From assistant to working partner

The initial expectation was faster access to data for managers, and İnci said that goal has been met. What became clear as the work went on was that the real change is more than seeing information only when it is asked for. The weekly performance report automatically includes store, channel, product and sales-consultant performance and anomalies, alongside sales and target data.

The change in perspective was summed up like this: AI is not only an assistant that answers questions; it can be a working partner that produces scheduled analysis, gets it to the relevant teams and makes unusual developments visible. For that to last, data quality and clear business rules are the basic condition.

On ease of use, İnci highlighted the approach to feedback: from day one the managers' daily questions were put at the centre, and which data they struggled to reach, which reports they wanted and in what format were worked through together. The product evolved with real user feedback.

What comes next?

İnci said they want to carry the current analysis and agent infrastructure into more operational use cases:

  • Finding the right product from a product photo, then matching outfits and similar products using photos and product data
  • Evaluating incoming e-invoices on ZZeti Zeka and posting them to Nebim V3 automatically
  • Allocation work that draws on stock and sales data

One piece of advice for those starting out

In the long run the expectation is for AI to become a helper that can start certain processes and produce recommendations. İnci also stated the basic approach plainly: the aim is not a system that makes uncontrolled decisions in place of people, but one that supports the decisions of authorised employees with verifiable data and business rules.

His advice to managers who want to start an AI transformation: before the model, be clear about the business problem you want to solve and about your data. Yargıcı did not try to solve everything at once; it moved step by step. Data preparation, integration, process design and end-user habits matter equally; when the right use case is chosen and business units are involved, adoption gets easier and the result has real value in daily work.

Takeaway

The Yargıcı example shows that enterprise AI sticks when it starts with one clear question and grows step by step: first natural-language access to data, then the channels employees already use, then agent automations. Underneath it all are data quality and clear business rules.

Frequently asked questions

What is Yargıcı AI?

It is the AI solution Yargıcı has built on ZZeti Zeka. Managers query sales, product, store and inventory data in natural language and reach it through WhatsApp and Teams, while agents prepare and send daily and weekly reports automatically.

What data does Yargıcı AI work on?

As described in the conversation: Nebim V3 data, the Nebim database and data warehouse, and Yargıcı's own data structures and business rules. Product photos and document content were also brought onto the platform.

Where is the recording?

In the "Türkiye Yapay Zeka Zirvesi - 2. Gün" stream on TRAI's YouTube channel. The video in this article opens at the moment the conversation starts. The conversation is in Turkish.