Skill-Based Culture: Three-Way Competency Assessment and Personalized

1. Create a single, objective competency assessment system for IT, Digital, Data and CISO employees.

2. Develop a unique competency matrix for each specialization, benchmarked against the market and tailored to the bank’s specifics.

3. Implement a three-way assessment based on a single matrix: self-assessment, AI-generated test and manager assessment.

4. Build an internal talent marketplace based on common criteria.

5. Gain visibility into every specialist’s actual skills and quickly assemble and reassemble teams around business needs.

6. Apply the same matrices in recruitment to compare external candidates directly with internal talent.

7. Identify skill gaps and strong competencies, find internal experts to act as trainers, and build personalized learning based on competency levels.

More than 1,000 employees in IT, Data, Digital and CISO were assessed between December 2025 and June 2026. Coverage is about 50% of Head Office headcount, excluding the contact center. • More than 170 specializations and more than 1,000 competencies have been defined. A bank of 250 questions has been created for each specialization (around 40,000 questions in total); the test lasts 90 minutes and every test is unique. • The assessment system makes it possible to evaluate whether an employee matches their current grade, identify skill gaps, and identify employees whose results exceed their current level.

Every employee receives personalized AI recommendations immediately after the assessment; the first learning paths and HR analytics are already live, and the full personalized learning module launches in November 2026. • The figures above reflect implementation and coverage results. The impact on business metrics, retention and productivity has not yet been quantified.

• Specialized AI platform for competency assessment - in-house development

• Generative AI: benchmarking matrices against the market, recommending competencies and weights, generating topics and test questions

• Three-way assessment module: self-assessment, AI test, manager assessment

• Results analytics: skill gaps and strong competencies, expert map

• Talent marketplace based on common matrices: searching for specialists by skill and assembling teams for specific tasks

• Use of the matrices in candidate recruitment

• In development: a learning platform based on assessment results and the matrices

• Manager engagement: the matrices are built by managers themselves, not by HR, so they had to be trained to work with the platform and AI and convinced that it is a management tool rather than a formality.

• Balancing AI and expertise: AI suggests competencies, weights and questions, but test quality depends on context and manager validation. A review process had to be built so that questions reflect the bank’s real specifics.

• Consistency vs. uniqueness: each specialization has its own matrix, but for the talent marketplace the criteria and scales must be comparable.

• Rapidly changing IT, Data and Information Security competencies: the matrices and benchmark need to be updated regularly.

• End-to-end logic: the same matrices connect assessment, recruitment, the talent marketplace and learning.

Role and skill requirements for IT staff  ( including AI) are changing very fast.

Halyk is not just a bank but an ecosystem in which new products and projects are launched continuously. The practice therefore addresses two key challenges: 1) Objective competency assessment that enables teams to be reassembled quickly. The business needs to understand each employee’s actual level and set of competencies in order to promptly form and reshape teams around new products and priorities, taking into account specialists’ levels and skills. 2) Clear development tracks for IT talent. Employees need to know where they can grow, see their strengths and gaps, have clear and transparent development paths within their specialization, and understand which skills they need to acquire or strengthen to move into a new role.

Assessment based on a common competency framework reduces reliance on unstructured, subjective judgments and makes development conversations between employees and managers concrete and data-driven. It makes it possible to target learning precisely at identified gaps rather than offering “the same for everyone”, and to find employees whose competencies exceed the requirements of their current role — candidates for internal experts, mentors and the talent pool. International practice confirms the value of this approach: according to Deloitte research, skills-based organizations are 107% more likely to place talent effectively and 98% more likely to retain high performers.

Halyk  assessed all IT staff.                                                                  

The competency were developed by business unit heads and subject-matter experts; HR provided a single methodology and supported the process. Managers took part in preparing the test materials at every stage, from building the matrices to validating the test questions.

Each employee’s assessment consists of three parts linked to the same matrix: 1) self-assessment — the employee rates their proficiency in each competency; 2) a 90-minute video-proctored knowledge test with 60 questions of varying difficulty; 3) the manager’s assessment against the same matrix.

To create valid tests, the state-of-the-art AI model Claude Sonnet 4.5 was used. Questions are generated by AI based on a reference description of the competencies and the role context. This made it possible to build an extensive question bank and ensure every test is unique.

Generating topics and questions was not left entirely to AI: validation remained with managers and experts. People decide what to test and how, and AI helps do it quickly and at scale.

The assessment results show competency proficiency levels, skill gaps, and employees with strong competencies who can act as internal experts.

Completing the assessment, employee gets personalized AI-generated recommendations based on results: strengths,...

Republic of Kazakhstan: IT, Digital, Data and CISO units.