Businesses across Europe and Central Asia are using artificial intelligence to reduce the time spent on routine operations, but the efficiency gains have yet to translate into measurable productivity improvements for most firms, according to a World Bank report.
Companies are deploying AI chatbots to handle customer enquiries, assist employees with internal information, support back-office operations and prepare documents. In some cases, tasks that previously took two weeks can now be completed within a few hours, the report said.
However, the World Bank’s October 2026 Europe and Central Asia Economic Update, titled Making AI Work: Jobs, Firms, and Productivity, found that these applications have largely reduced operational time without producing measurable effects on overall productivity.
The findings highlight a challenge facing businesses seeking returns from AI investment. Adopting the technology and completing individual tasks faster do not necessarily mean a company is producing more output, generating higher revenue or improving its overall performance.
The report said a relatively small group of firms has moved beyond basic applications to integrate AI into core business operations, including financial institutions using the technology to assess loan applications, retailers making stock replenishment decisions and energy companies optimising production inputs.
These applications represent a deeper use of AI because they influence operational decisions rather than simply assist employees with routine tasks.
Measuring the returns
The World Bank said measuring AI’s contribution to productivity remains difficult, even for businesses that report improvements in efficiency.
For instance, an internal AI chatbot may reduce the time employees spend answering routine questions, but the resulting productivity gains can be difficult to quantify across the entire company.
Companies using AI in customer-facing services also struggle to determine whether revenue generated through the technology represents additional business or merely replaces sales that would have occurred through existing channels.
The report noted that AI has also enabled businesses to undertake tasks and processes that did not previously exist, making it difficult to establish what their performance would have been without the technology.
This uncertainty extends to companies developing AI solutions, which also face difficulties quantifying returns on investment, a key consideration when deciding whether to expand projects beyond initial testing.
Management and data constrain adoption
The report identified management capabilities and organisational readiness as major factors determining whether businesses can turn AI adoption into economic gains.
Companies need managers who can identify processes where AI can add value, redesign workflows around the technology and guide employees through organisational changes.
The availability of AI tools alone is insufficient, the World Bank said, pointing to the shortage of business-oriented engineers and managers capable of connecting technical solutions to commercial needs.
Data readiness presents another constraint. Although businesses generate substantial amounts of information, much of it remains stored in separate systems or in formats that are difficult to use.
The report said the main challenge for many companies is not access to AI models but the availability of usable, integrated data.
Concerns about data leakage, the accuracy of AI-generated outputs, uncertain financial returns and changing regulations are also discouraging businesses from moving beyond experimental applications.
A small group moves into core operations
The report cited examples of businesses that have integrated AI more deeply into their operations, illustrating the potential gains when the technology is linked to specific commercial objectives.
A Serbian retail company used an AI call-centre agent to reduce call volumes by 23% in its first month and almost 50% after three months.
In Kazakhstan, a bank automated almost 100% of unsecured consumer-loan credit decisions, allowing funds to reach customers in slightly more than a minute.
A financial institution in Serbia also deployed 20 AI agents to perform repetitive transfers and data-extraction tasks, with the tools estimated to save about 20,000 hours of manual work.
In Croatia, an insurer’s fully digitalised claims process handled thousands of claims in a single day during a natural disaster, work that previously required months.
These examples show how AI can affect processing capacity, service delivery and operational decisions when businesses integrate it into their core processes.
The World Bank said the firms achieving more extensive adoption tend to have addressed constraints around workforce skills and data integration, with management focused on investment returns rather than adopting AI for its own sake.
The report argues that the wider economic benefits of AI will depend on businesses moving beyond isolated efficiency improvements to reorganise operations around the technology.
For companies across the region, the challenge is therefore shifting from acquiring AI tools to demonstrating that their use can deliver sustained improvements in productivity and business performance.
