DNB plans a reduction of around 400 full-time equivalents — Latvia will also be affected

DNB's 6 October 2026 plan covers a reduction of around 400 FTEs across Technology & Services.

What is known about Latvia?

In LTV's 24 October report, a bank representative confirms Latvia is included, without specifying numbers. An anonymous employee describes around 175 people in Riga facing uncertainty, not 175 dismissals. Employees have been promised information about the next steps on 1 November.

Eurostat defines FTE as a unit comparing employment through working hours. A number of full-time equivalents should therefore not automatically be converted into a count of individual dismissals.

Is AI the whole explanation?

DNB identifies AI uses in customer-data controls, Know Your Customer work and coding; it is also simplifying processes.

A 2025 International Labour Organization study assesses occupational tasks' exposure to generative AI. It considers job transformation the most likely overall effect because many tasks still require human input. That study does not determine any particular DNB employee's future.

Editorial conclusion: distinguish technology's ability to perform a task, a company's decision to reorganise work and an individual role's future. Software may accelerate one stage while another still needs verification, exception handling or customer communication.

What should employers check before making changes?

The following is an editorial assessment plan, not a description of DNB's internal procedures:

Describe work at task level. Identify incoming data, the required output and cases needing individual judgement. A job title alone does not reveal which work can be automated.
Compare the complete process. In a trial, assess time spent on checking, corrections and follow-up communication as well as producing the initial result. Set the same quality criterion for both ways of working.
Explain how saved time will be used. Is the aim to handle more volume, reduce backlogs or redistribute duties? Each option needs a different staffing and training plan. Show tool maintenance and output monitoring separately in savings calculations so an initially attractive speed comparison does not obscure them.
Assign responsibility for exceptions. Record who handles complex cases, who can stop a faulty process and how changes reach the team. Identify the responsible person and clear cover arrangements for their absence.

What should employees ask management?

Prepare specific questions: which duties will change, what criteria will determine the skills needed, are alternative roles or training planned, and when will individual information be provided? Ask which decisions are final and which remain under discussion.

Afterwards, send a short summary of your understanding and invite corrections. Note the contact person and next discussion date. This helps distinguish an answer received from a question still unresolved.

Create a short map of your tasks. For each, identify where industry knowledge, checking results or cooperation with another team is needed. This enables discussion of a specific development need instead of a general promise to learn AI.

Before choosing a course, agree on its intended practical outcome: for example, independently checking an automated process's output or documenting an exception. Discuss working time allocated to learning so the training idea becomes feasible.

For legal and administrative changes within a company, consider services for businesses. To compare external business service offers, explore the listings section, checking each provider's competence and terms individually.

Official information sources

Additional news source

LTV/LSM, 24 October 2026