The share of companies in Türkiye using artificial intelligence (AI) has roughly doubled in a year. According to the bulletin that TÜİK (the Turkish Statistical Institute) published on October 2, 2026, among enterprises with 10 or more employees this rate rose to 14.0% in 2026, up from 7.5% in 2025. As usage spreads, the question facing executives is changing too: "Should we use AI?" is giving way to "How should we acquire this solution?"
TÜİK's 2026 bulletin sheds light on this question with data for the first time. A heading that did not appear in the previous year's bulletin has been added: the routes by which AI-using enterprises acquire this software and these systems. In this article we look at that data, at differences by company size, and at the personal-data dimension.
How do companies acquire AI?
TÜİK asked enterprises that use AI how they acquired their solutions. According to the bulletin, the route that stands out is ready-made software:
- Using open-source software (free or paid): 67.8%
- Using closed-source software (free or paid): 49.1%
- Having software or systems developed by external providers: 46.5%
- Developing software or systems with their own employees: 32.1%
The percentages add up to well over 100. This shows that a significant share of enterprises do not choose a single route but use several together. For example, the open-source and closed-source shares alone add up to 116.9%; so at least one in six AI-using enterprises uses both types of ready-made software together.
The table tells us three things. First, the starting point is mostly ready-made tools. Second, having an external provider develop the solution is far from a marginal route: roughly one in every two AI-using enterprises uses this route as well. Third, although in-house development is the least-used method, close to a third of AI-using enterprises also run this capability inside the organization.
Differences by size: large companies are far ahead
The choice of acquisition route cannot be considered independently of company size. According to TÜİK data, the share of enterprises using AI in 2026 is 12.8% among those with 10–49 employees, 17.0% among those with 50–249 employees, and 37.1% among those with 250 or more. The same rates in 2021 were 2.3%, 3.6% and 9.6%, respectively.
Compared with the European Union, the gap is still marked. According to Eurostat, 19.95% of enterprises in the EU used AI technologies in 2025; among large enterprises the rate was 55.03%. Because the two institutions report different years and different samples, the comparison should be read as indicative.
There is a size gap among companies considering adoption as well. The share of enterprises that do not use AI but are considering it is 8.3% overall in 2026; it is 7.4% among those with 10–49 employees, 12.2% among those with 50–249 employees, and 14.9% among those with 250 or more.
The TÜİK bulletin does not break down acquisition routes by size group. A conclusion such as "large companies develop more in-house" therefore cannot be drawn from this data. What can be said is this: among large companies, use is both more widespread and growing faster, while mid-sized companies are markedly ahead of small ones in their inclination to consider adoption.
The personal-data dimension: one in every six users
The acquisition route also determines where, and under whose control, data is processed. According to TÜİK, 16.6% of AI-using enterprises declared that they process personal data with these technologies, including gender, age, occupation, address, purchase records or facial images. This share is barely affected by scale: 16.9% among enterprises with 10–49 employees, 15.3% among those with 50–249 employees, and 17.0% among those with 250 or more.
This point is directly tied to the acquisition decision. According to a guide published in November 2025 by the Kişisel Verileri Koruma Kurumu (KVKK, Türkiye's Personal Data Protection Authority), a developer that processes personal data on the instructions of a third party, within a framework whose purposes and means are determined by that party, can be characterized as a data processor. The guide also stresses that roles are determined not by the wording of contracts but by the parties' actual control. In practice, this means that in a system developed by an external provider, the company must clearly define the decisions on why, from what source and for how long data will be processed. This assessment reflects what the source says and does not replace legal advice.
Strengths and weaknesses of the three routes
The framework below contains no source figures; it is a thinking tool proposed for turning the table in the TÜİK data into a corporate decision-making process.
Ready-made software (open or closed source). The fastest way to get started. Generally sufficient for common needs such as text summarization, classification or document reading. Open-source solutions can offer advantages in flexibility and in keeping data inside the organization; however, responsibility for installation, updates and security remains with the organization. With closed-source solutions this burden is lighter; in return, visibility into how the model works and how data is processed may be limited.
Commissioning an external provider. The way to reach a customized solution when there is no in-house expertise. The key to success is for the organization to define the need and the success criterion before the provider does. Maintenance, model updates and who will operate the system after delivery should be settled at the contracting stage; otherwise the organization may become dependent on a system it does not understand.
Building with your own team. The route that offers the most control but demands the longest-term commitment. It makes sense for use cases that create a competitive advantage and rely on organization-specific data. However, it requires data infrastructure, skilled staff and continuous maintenance; if it will remain limited to a single project, the investment generally struggles to pay off.
Five questions for the decision
- How standard is the use case? If many companies in the market have the same need, a ready-made solution is probably sufficient.
- Does the solution create a competitive advantage? If it rests on organization-specific processes and data, custom development comes into play.
- Will personal data be processed? If so, the data flow, retention period and the roles of the parties should be put in writing from the outset.
- Who will take ownership after delivery? Even when working with an external provider, there should be someone inside the organization who understands the system.
- Can the routes be combined? The TÜİK data shows that companies use the routes together; customizing a ready-made model with external support and growing the internal team over time can be a common transition model.
Conclusion
TÜİK's 2026 data shows that AI in Türkiye is acquired mainly through ready-made software, but that having an external provider develop it and in-house development also hold an important place. The right path is not a single choice but a mix that varies with the use case, the sensitivity of the data and the organization's capability.
From Orient Global's perspective, separating use cases into "standard" and "organization-specific" before the acquisition decision is a good starting point. Over the long run, dependency is reduced by moving quickly with a ready-made solution for standard needs and, for organization-specific areas, building from the first project onward a structure that keeps the accumulated know-how inside the organization. In every scenario involving personal data, clarifying the parties' roles up front also prevents uncertainty later on.
Why a logistics lens helps here. Choosing between ready-made software, an external provider and an in-house team is, at heart, a sourcing decision, and sourcing is everyday work in operations and supply chain: buy standard items off the shelf, source specialized ones from a qualified supplier under clear terms and service levels, and keep in-house whatever sets the business apart. One of Orient Global's consultants brings more than 30 years in operations, logistics and supply chain, including running logistics, warehousing, technical service and administrative operations end to end and building and leading teams of 100+ people. An industrial engineer, he now brings that discipline to SMEs through AI and automation. Meet our consultants.
#interimmanagement #forwarddeployedengineering