The Future of Software Development in T24 is Here

Josef Mayrhofer

6 hours ago

3 min read

Coding in the T24 Globus release started with the INFOBASIC programming language. Almost all core routines/applications are developed using this language.

T24 later evolved from GLOBUS to TAFC and TAFJ. During this transition, the coding changed from INFOBASIC to Componentized, and then to Java. The extensibility framework also uses Python coding now. Depending on the T24 release and framework, banks can choose the required programming language.

Traditional Development and Deployment in T24:

The usual way of coding in T24 involves preparing the technical specification document from the business, getting it reviewed, writing the code manually, doing the unit testing, and sending the code and test reports for approval.
Once the lead approves the code, it enters the testing phase; when it passes all testing in environments like SIT, UAT, and PREPROD, the change is approved in a meeting with all the testing evidence.
Processes like these are time-consuming and resource-intensive. For this reason, Temenos has introduced the AI-powered Software Development Lifecycle (SDLC). Most of the activities are automated and AI-driven, with minimal human support.

The New AI-Powered SDLC of Temenos:

Temenos has adopted Anthropic AI for its software development
Adopting AI can improve scalability, compliance, regulatory requirements, and traceability, which reduces delays caused by communication gaps.
This approach creates a multistep workflow across the entire SDLC. It is aware of the codebase, project plans, architectural patterns, and connected tools, while also generating test cases and reviewing code. AI supports every role, not just developers, while requiring minimal human involvement.
Claude Code AI coding agent is live now, and a phased rollout for spec-driven SDLC is considered. Backpatching for the previous release is also considered.

Key Features

AI as a Business Analyst:

The business requirements are analyzed from the BRD and translated into the required technical documentation. The AI also handles coding, test case generation, and test execution.

Other Roles and Responsibilities:

Developer: Writes and modifies code, handles T24 requests, and implements technical requirements.

Code Reviewer: Reviews the developer's code, checks for correctness, and manages pull requests via Git.

Release Manager: Responsible for deploying tested code across all environments and creating hotfix branches.

Branching Models

Trunk-Based Development (TBD): Developers work on a common trunk branch and use trunk branches in order to achieve continuous integration and delivery by frequently merging their changes.

Release vs Hotfix Branching: Developers use development branches, while hotfix branches are generated for urgent corrections, along with relevant pull requests and release package builds.

Release Management

Release Manager generates the packages from Workbench and upload it to artifact repositories. The deployments in T24 usually have the following sequence: DEV>SIT>UAT>PRE-PROD>PROD.

Testing Management

Using ATF (Automated Test Framework), test cases are executed through scripts containing OFS details across all environments. The Karate test framework can be used to develop RVTs (Release Validation Tests), ensuring all functionalities are covered before being included in release packages.

DevOps and Continuous Integration

The Software Development Life Cycle (SDLC) supports continuous integration and delivery through regular commits, automated builds, and testing, reducing merge conflicts and improving codebase stability.

Tenemos Copilot:

AI Copilot is an AI-powered assistant embedded directly inside Temenos core, designed to make core banking users more productive.
Problems AI Copilot can solve include: Customer-related info, instructions on bank processes, support with errors, monitoring batch performance, follow-ups on expiring arrangements, insights on existing data, initial test cases, etc.
Conceptual architecture containing the major components like conversation and completion services that handle the AI copilot.

tGuru as a coding assistant:

tGuru is an AI-powered chatbot that answers questions about T24 products, generates scripts and code, and identifies errors.
Using tGuru throughout the entire development and testing process is not yet recommended. However, its knowledge of T24 documentation can help clarify requirements and identify gaps or flaws.
tGuru can also help identify performance, functional, and coding issues and provide recommendations for improvement.

Implementing AI in T24 will significantly benefit both developers and bank employees. AI, along with human support in core banking, can lead to more efficient implementations, especially during upgrade development and testing. Additionally, bank operators can utilize the AI Copilot to assist with their daily tasks. Ultimately, a human-AI integrated system simplifies the validation and design of software, particularly within core banking like T24.