Digital Program Evaluation in Eurasia: How AI and Open Data Are Transforming the Development Approach

What is Impact Evaluation and Why Do We Need It?

When governments, large corporations, charitable foundations, or international organizations launch social or environmental programs and projects—building new schools, implementing financial literacy programs, or introducing air pollution control measures—the question always arises: did this actually deliver real benefits?

Typically, effectiveness is assessed through formal reports and direct outputs: how much money was spent, how many lectures were delivered, and how many people received certificates. But this doesn't answer the fundamental question: Have people's lives actually improved? Have their opportunities expanded? These are precisely the questions that impact evaluation answers.

Impact evaluation is not simply a financial audit or review of reporting documents. It's an in-depth study that helps us understand the value of various interventions and distinguish the real long-term effects of a project or program from random external factors. Evaluation allows us to understand what specifically worked, why it happened, and how to make future projects even more effective while eliminating wasted resources. In the context of digitalization, this work becomes especially significant: data volumes are growing, decisions are made faster, and consequently, the need increases for tools that help distinguish formal indicators from actual results.

Global Context: gLOCAL Evaluation Week 2026

These very questions were the focus of gLOCAL Evaluation Week 2026—an annual gLOCAL evaluation week that brings together monitoring and evaluation specialists, representatives from government agencies, business, NGOs, the academic community, and international organizations. In 2026, it took place from June 1-5 and was dedicated to the theme of trust, evidence, and evaluation in the age of artificial intelligence. Throughout the week, professional associations and research organizations from various countries held their own events, including sessions for the Russian-speaking evaluation community and the Eurasian Alliance of National Evaluation Associations.

At the Eurasian Evaluation Alliance webinar, experts from Russia, Tajikistan, Kazakhstan, Armenia, and Kyrgyzstan discussed how digitalization, artificial intelligence, and new data requirements are transforming the evaluation of government, social, and corporate programs. The event was moderated by Konstantin Mishenichev—Chairman of the Eurasian Alliance of National Evaluation Associations and Associate Professor at Almaty Management University. 

"Against the backdrop of information space fragmentation in the Eurasian region, it's especially important to maintain professional connections among impact evaluation specialists, exchange experiences, and jointly develop approaches to the responsible use of data and AI," noted Konstantin Mishenichev, Chairman of the Eurasian Alliance of National Evaluation Associations.

Transformation of Government Program Evaluation in Tajikistan

One of the most comprehensive cases was presented by Khayriniso Rasulova, an evaluation expert from the Republic of Tajikistan. In late April 2026, the country held its first international conference on monitoring and evaluation, attended by approximately 470 representatives from government agencies, the academic community, civil society, the private sector, and international organizations. The conference resulted in the adoption of a national declaration and preparation of a roadmap for building a monitoring and evaluation ecosystem. The document includes 12 areas—from legislative and methodological frameworks to financing, digital transformation, evaluation culture, and strengthening the data collection system.

"For us, it's important to build a culture of monitoring and evaluation (M&E)—to change the perception of evaluation specialists from auditors or inspectors to experts who help implement strategic documents more effectively. It's precisely this mature M&E culture that creates a favorable environment for building clear analytical algorithms. This is a fundamentally important stage, because digitizing inefficient processes only scales up their systemic flaws—you can't digitize chaos and expect to get order. Only when the algorithms are refined and the entire system begins functioning smoothly without digitization can we move toward its digital implementation. The result will be a fully functional monitoring ecosystem. This entire process cannot be the task of a single ministry: it requires collaboration among government, the academic community, civil society, the private sector, and development partners," noted Khayriniso Rasulova, coordinator of the monitoring and evaluation community of practice in Tajikistan and board member of the Public Association "Evaluation for Sustainable Development." 

Special emphasis was placed on personnel training. With support from JICA experts (Japan International Cooperation Agency), the Secretariat of the National Development Council under the President of the Republic of Tajikistan/Ministry of Economic Development and Trade of the Republic of Tajikistan developed for the first time a "Guide to Formulation, Monitoring and Evaluation of Strategic Planning Documents in the Republic of Tajikistan" as a methodological document based on the theory of change approach. Based on this guide, training-of-trainers sessions were conducted at the Academy of Public Administration with assistance from the Civil Service Agency of the Republic of Tajikistan, and for faculty at leading universities in Tajikistan with support from the Ministry of Education and Science of the Republic of Tajikistan. 

Also based on this Guide, 140 civil servants from 68 cities and districts in Tajikistan have already been trained, and starting in 2027, a monitoring and evaluation course is planned for inclusion in the Academy of Public Administration curriculum. For the region, this could serve as an example of transitioning from fragmented evaluation practices to an institutional system in which government, academia, civil society, and international partners work within a unified framework.

Digitization of evaluation in Kazakhstan: expanding the open data base and implementing AI in social project evaluation practice

Open data for long-term impact evaluation

Kazakhstan's experience demonstrates another important trend—expanding the very database for evaluation. According to Aliya Sarsekeyeva, project director of Kazakhstan Sociology Lab, digitization is changing the logic of social research: administrative data, digital footprints, job vacancy data, educational platform data, and urban infrastructure data are being added to traditional surveys. 

In the research sphere, Kazakhstan Sociology Lab studies numerous topics: from data on student academic performance and olympiad participation to the influence of urban environment on academic achievement. To accomplish this, modern tools of advanced statistical and network analysis are being actively implemented, as well as spatial analysis using geographic information systems and geo-data. This approach expands evaluation capabilities: digital data allows us to look not only at the fulfillment of program indicators, but also at long-term effects, territorial differences, and structural factors of inequality.

"Digitization is changing the very logic of social research. It's not about abandoning surveys or other classical methods, but about the data ecosystem becoming broader: administrative data, digital footprints, labor market data, education data, and urban environment data are being added to it. This allows us not only to capture individual indicators, but to understand more deeply how social processes work and what factors influence program outcomes," noted Aliya Sarsekeyeva, project director of Kazakhstan Sociology Lab.

AI-based chatbots for project evaluation 

Another practical case of AI application in evaluation was presented by Maria Lokteva, head of impact research and evaluation at inDrive. The company operates in more than 45 countries worldwide and uses artificial intelligence not only to evaluate business metrics, but also for research related to social impact and the design of corporate social responsibility programs. One case study involved examining sensitive social norms in Kazakhstan related to the concept of "uyat"—reputation, shame, family pressure, and gender expectations.

As part of the study, 50 interviews were conducted using an AI chatbot. This format helped lower the barrier to discussing sensitive topics: respondents could answer in a safer and less socially risky environment, including through voice messages. AI helped collect and transcribe responses, code materials, identify recurring themes, contradictions, and differences between generations and gender groups. 

This case demonstrates that AI can become not just a tool for accelerating analytics, but also a way to obtain data on topics where traditional methods face high social barriers. For corporate foundations, ESG teams, and organizations implementing corporate social responsibility programs, leveraging the infrastructure of major tech corporations opens broad opportunities for adopting digital tools to assess the impact of programs and projects. 

AI in Government Regulation in Kazakhstan

A broader framework for applying AI in public administration was proposed by Vadim Novikov, advisor to the president and professor at Almaty Management University. He distinguished between two areas of AI application: political decisions, which involve values, rights, risks, and cost distribution, and administrative-analytical tasks, where AI can help the state see the consequences of its own decisions more quickly. 

Professor Novikov proposed distinguishing between a "listening state" and a "sensing state." The former collects signals from citizens—appeals, complaints, comments, consultation results. The latter learns from the consequences of its own decisions. This framework holds particular significance for regulatory policy. AI can help analyze thousands of comments on draft regulations, group business complaints, identify recurring barriers, compare document versions, detect regulatory duplication, and track how rules work after adoption—through complaints, courts, inspections, procurement, and enforcement. 

In Armenia, Evaluators Actively Use Modern AI Agents at All Stages of Research, Monitoring, and Evaluation

Practical experience using AI in monitoring and evaluation was presented by Kristina Ter-Abarmyan, research director at Prisma agency in Armenia. According to her, AI agents such as ChatGPT and Claude have already become everyday tools for preparing proposals, developing samples, indicators, theories of change, data collection instruments, transcribing interviews, and preliminary analysis of materials. For specific tasks, such as sample development, AI reduces work from one or two days to several hours.

"AI doesn't replace the expert. Its role is to accelerate routine operations, help structure information, and prepare first drafts of documents. The new professional competencies for evaluators include the ability to formulate quality queries, critically verify AI results, understand model limitations, and comply with data confidentiality requirements," said Kristina Ter-Abramyan, research director at Prisma agency.

Data protection became one of the key discussion topics. In research teams' practice when working with AI, they use transcript anonymization, removal of identifying information, and prior coordination of AI tool usage with clients and donors. This is especially important for evaluating social programs, where researchers often work with sensitive information about beneficiaries, vulnerable groups, and personal stories.

Assessing Social Effects in the Context of Russia's Transition to Digital Sovereignty

The Russian context was presented through the lens of digital sovereignty. For evaluation specialists, this means simultaneously expanding government digital systems while complicating access to some foreign services, including certain AI tools. This stimulates the development of national platforms, domestic software, and new requirements for personal data processing, but also creates risks of data fragmentation and limitations on international professional exchange.

"Digitalization of evaluation in the Eurasian region is developing unevenly, but is already becoming a stable trend. For the state, it's an opportunity to assess program effectiveness more accurately and adjust policy more quickly. For business and corporate foundations, it's a tool for more transparent assessment of social impact and ESG projects. For the research community, it's a transition to new methods where classical expertise combines with big data analysis, AI, and digital platforms," said Konstantin Mishenichev, chairman of the Eurasian Alliance of National Evaluation Associations.