A featured contribution from Leadership Perspectives, a curated forum for enterprise technology leaders, nominated by our subscribers and vetted by the CIOApplications Editorial Board.

Applus+ IDIADA

Software Architecture and Web Development Involving GenAI and Coding Assistants for Corporate Development Teams in 2025

In today’s fast-paced software engineering landscape, corporate development teams must deliver secure, scalable and user-centric solutions while boosting productivity. At Applus+ IDIADA, our Architecture team—responsible for security, architecture and UI/UX—has adopted Generative AI (GenAI) tools to transform how we design, develop, and deploy software. All our software teams actively use these tools. We’ve integrated Amazon Q Developer Pro as our primary AI assistant, securely connected to our on-premises GitLab, and enhanced it with AIDA, our proprietary AI chatbot. This setup has streamlined legacy modernization, testing and cross-team collaboration. Developers primarily use Visual Studio Code and IntelliJ Ultimate, with GenAI embedded in both, enabling productivity gains without workflow disruption—leveraging a Java 17, Angular 18 and Python 3.12 stack.

The Rise of GenAI in Corporate Software Development

By 2025, GenAI has transitioned from a novelty to a cornerstone of enterprise software development. These tools are no longer just “code completers” but are collaborative partners in architecture design, security auditing and legacy system modernization. For teams like ours, by balancing cross-functional responsibilities, GenAI bridges gaps between disciplines, automates repetitive tasks and surfaces insights that human developers might overlook—all while adhering to strict corporate governance frameworks.

  • GenAI is no longer just a coding assistant. It is a trusted partner in design, security and collaboration. At Applus+ IDIADA, our integrated approach shows that real innovation comes from tools that adapt to your environment and your people


Our Toolbox: Integrated AI for Enterprise-Grade Development

Amazon Q Developer Pro: The Secure, Context-Aware Backbone

Amazon Q Developer Pro is central to our AI-driven workflow and is integrated directly into Visual Studio Code and IntelliJ Ultimate. It operates in the full context of our on-premises GitLab repositories, including commits, merge requests and pipelines. This enables automated Java 8 to 17 migrations, refactoring deprecated APIs and generating documentation per pull request. Unlike GitHub Copilot, which posed privacy issues, Amazon Q functions securely within our environment.

A recent Angular upgrade resolved 92 percent of dependency conflicts while ensuring compliance with internal security policies. We also use its dashboards for real-time insights into productivity, code quality and AI adoption. Developers using Amazon Q in IntelliJ reduced code review cycles by 30 percent compared to manual workflows.

AIDA: Bridging Teams and Tools

Artificial intelligence digital assistance (AIDA), our proprietary chatbot, acts as a unified interface for cross-tool collaboration. Hosted on AWS and integrated with different corporate tools, AIDA answers developer queries by synthesizing data from GitLab, Jira and Amazon Q. For instance, a developer might ask, “AIDA, how many Java 17 migrations are pending for the vehicle diagnostics module?” and receive a response combining GitLab issue statuses, Amazon Q migration logs and historical sprint velocities.
During a critical incident involving a Python data pipeline failure, AIDA correlated error logs from GitLab CI/CD, AWS CloudWatch metrics and Microsoft Teams chat histories to identify a misconfigured environment variable within minutes. It then drafted a post-mortem report, which Amazon Q later enriched with code snippets to prevent recurrence.

Evaluating GitHub Copilot, Replit and Cursor under Privacy Limits

While we conducted a proof of concept (PoC) with GitHub Copilot and found its code suggestion capabilities impressive, its inability to integrate with our on-premises GitLab instance—due to OUR data privacy regulations—limited its utility. Unlike Amazon Q, which operates within our secure perimeter, Copilot’s cloudcentric model conflicted with our requirement for complete codebase isolation. Nevertheless, the PoC highlighted the potential of AI-assisted prototyping, which informed our stricter governance policies for Amazon Q.

We are evaluating Replit for cloud-based Python scripting environments and Cursor for its AI-assisted debugging in IntelliJ. Early tests suggest Cursor could reduce Java concurrency bug resolution times by 40 percent, though we remain cautious about data residency requirements.

Key Benefits for Security, Architecture and UI/UX Security

Amazon Q’s GitLab integration ensures security by default, automatically redacting secrets from code commits and enforcing IAM policies aligned with our on-premises infrastructure. During a recent audit, it detected an attempt to commit a hardcoded AWS key, blocking the merge and alerting the team - all within seconds.

Architecture

The combination of Amazon Q and IntelliJ Ultimate has transformed how we design systems. For a high-throughput telemetry processing project, Amazon Q proposed a Kafka-based event-driven architecture after analyzing similar patterns in our GitLab history. At the same time, AIDA cross-referenced the design with past performance reports to recommend optimizations.

UI/UX

Angular developers using VS Code benefit from Amazon Q, developing quick UI mockups with Angular 18 components for quick validation with customers.

Challenges and Best Practices

1. IDE Plugin Fatigue: While Amazon Q’s IntelliJ and VS Code plugins are powerful, some developers initially struggled with context-switching. We addressed this by creating custom AI shortcut profiles.

2. Privacy vs. Innovation: The GitHub Copilot PoC reinforced the need for tools that balance AI capabilities with on-premises data control—a balance Amazon Q achieves.

3. AI Trust Calibration: Early overreliance on Amazon Q’s suggestions led to a few flawed migrations. We now enforce “AI-assisted, human-validated” workflows.

GenAI’s Future as Architect and Team Partner

We’re expanding Amazon Q’s role to auto-generate RFCs for architectural changes and self-optimize GitLab pipelines based on historical success rates. Meanwhile, AIDA is evolving into a developer mentor, offering personalized upskilling plans based on GitLab commit patterns and code review feedback.

Our vision is for GenAI to become a copilot and an autonomous systems architect—refactoring, documenting and even simulating architectural decisions in real-time.”
At Applus+ IDIADA, GenAI isn’t just a tool—it’s a contextualized, compliant extension of our team. By embedding Amazon Q into VS Code and IntelliJ, governing it through GitLab integration and augmenting it with AIDA’s cross-platform intelligence, we’ve reduced technical debt and accelerated feature delivery. For enterprises navigating the AI uprising, our experience underscores a critical lesson: success lies in tools that adapt to your ecosystem, not the other way around.

The articles from these contributors are based on their personal expertise and viewpoints, and do not necessarily reflect the opinions of their employers or affiliated organizations.