Summary

• Cognizant will support the EBRD's Tech Blueprint programme through outcome-led managed services and fixed-capacity services.
• The five-year framework covers workplace technology, cloud, data, security, software engineering, automation and AI.
• Cognizant proposes three AI products for technology operations, software development and employee support, but their deployment and financial benefits remain prospective.

The European Bank for Reconstruction and Development has selected Cognizant as a strategic technology partner under a five-year framework covering managed services, cloud infrastructure, software engineering, cybersecurity and artificial intelligence.

The agreement, announced on 26 October, will support the EBRD’s Tech Blueprint programme, which aims to modernise the systems and operating processes supporting the bank’s international activities.

Cognizant intends to deliver services through two complementary commercial arrangements: outcome-led managed services and fixed-capacity services. The framework also includes a programme of continuous improvement, with proposed AI tools intended to support technology operations, software development and employee services.

Neither organisation has disclosed the value of the framework, the distribution of work between its delivery models or the performance targets against which the proposed improvements will be measured. Those details will influence whether the arrangement can deliver efficiencies beyond the transfer of existing activities to an external supplier.

Two models divide operational responsibilities

Outcome-led managed services and fixed-capacity arrangements establish different expectations for the organisations delivering and receiving technology support.

Under an outcome-led model, a supplier can be responsible for achieving defined service results rather than simply providing specified staff or equipment. The contractual measures may concern reliability, responsiveness or other operational characteristics, although the EBRD has not disclosed the detailed outcomes applicable to this framework.

Fixed-capacity services, by contrast, provide an agreed allocation of people or specialist resources that can be directed towards particular activities. Such arrangements can be useful where an organisation needs additional engineering capability but retains direct control over priorities and delivery.

Combining the approaches may allow the bank to maintain routine technology services while allocating additional expertise to modernisation projects. However, the relationship between ongoing operations and new development must be governed carefully.

Where the same supplier supports an existing application and participates in replacing it, responsibilities for maintaining service quality, prioritising investment and managing transition risks can overlap.

The customer also needs sufficient internal technical knowledge to challenge supplier decisions and evaluate whether proposed changes offer value. Outsourcing activities does not remove accountability for technology governance, particularly within an institution operating across multiple jurisdictions.

Three AI products enter the delivery plan

Cognizant plans to contribute Neuro IT Operations, Flowsource and WorkNEXT as part of the wider programme. The products address different requirements across the bank’s technology environment.

Neuro IT Operations is designed to improve visibility across systems and help identify or resolve incidents more effectively. Its proposed capabilities include AI-assisted analysis intended to strengthen service resilience.

Flowsource applies generative and agentic AI to software engineering, with functions intended to support development, testing and earlier identification of problems. Cognizant also describes self-healing capabilities within the product, although it has not specified which activities the EBRD intends to automate or how much authority those systems will receive.

WorkNEXT addresses workplace technology services, including personalised support, voice and translation capabilities, and the coordination of employee requests through AI-assisted processes.

These technologies could reduce repetitive work across established IT functions, but their introduction requires accurate information about the systems on which they operate.

An incident analysis tool, for example, must distinguish an actual infrastructure problem from misleading monitoring information. If it recommends changes based on an incomplete understanding of system dependencies, an attempt to restore one service could create disruption elsewhere.

Software engineering agents face comparable constraints. Generated code must be tested, reviewed and integrated with existing development procedures, while automated changes to production environments require clearly defined approval arrangements.

Employee support systems also encounter information that may be sensitive, including identity records, access permissions and details of internal technology services.

The company’s plans therefore require more than installing three software products. Their integration must reflect the EBRD’s security requirements, information architecture and established operational responsibilities.

Institutional governance shapes implementation

The EBRD is a multilateral development bank operating in 40 economies across three continents. Its technology infrastructure supports investment activity, financial operations, reporting and the internal processes required to manage programmes across different jurisdictions.

That environment creates demands for continuity, accountability and information governance that differ from those associated with a limited office productivity deployment.

A service interruption affecting a central application could have consequences for multiple teams and locations, while changes to information systems may require careful consideration of regulatory and institutional requirements.

External technology providers can supply specialist expertise and help maintain complex systems, but the bank must remain able to understand the services on which it depends.

That includes access to documentation, operational records and sufficient technical knowledge to manage incidents or change suppliers if necessary.

The introduction of AI brings an additional requirement to establish responsibility for automated actions. Monitoring software may produce recommendations that are reviewed by engineers, while other tools could eventually undertake defined tasks without direct human involvement.

Those operating arrangements need to specify who authorises changes, which actions are permitted and how unexpected outcomes will be investigated.

Supplier contracts can define responsibilities, but accountability also depends on the ability to reconstruct events from reliable technical records. Where several systems exchange information, auditability can become more difficult unless records are designed to connect decisions with their underlying inputs.

Measuring value requires a baseline

The five-year framework gives the organisations a period in which to modernise services and measure performance against agreed expectations.

However, improvements in technology operations can be difficult to evaluate when the scope of services changes during the contract. A reduction in incident numbers, for example, may reflect better infrastructure management, altered reporting practices or the removal of an unreliable application.

Likewise, faster software development does not necessarily improve service quality if it produces additional maintenance requirements or defects.

Measures of performance need to account for reliability, security, customer and employee experience, and the cost of supporting technology throughout its operating life.

AI introduces another consideration because some activities may become less expensive to perform while creating new expenditure on computing, integration and supervision.

An outcome-led agreement may provide incentives for the supplier to automate appropriate work, but the financial treatment of those efficiencies will depend on the commercial terms.

The organisations have described their intention to improve operational efficiency and strengthen access to specialist capabilities, although they have not published baselines allowing those expected benefits to be quantified.

Cognizant’s selection provides a delivery partner for the EBRD’s technology modernisation programme. The evidence of success will emerge through completed implementations, service performance and the bank’s ability to retain effective control over increasingly automated operations during the five-year relationship.