Systems | Development | Analytics | API | Testing

When a user asks your AI assistant a question, who actually runs the database query? In most enterprise AI deployments, the answer is troubling: a shared service account with broad access to everything. The user's identity evaporates the moment their request enters the AI system. This architectural pattern creates security gaps, compliance failures, and data leakage risks that undermine enterprise AI adoption.

AI is changing how work gets done. But for many data leaders, it’s also creating a new challenge: managing the cloud bill. As more people (and more AI agents) query data, cloud data warehouse (CDW) spend can spike fast. Costs become harder to predict, and teams end up making tradeoffs—scaling AI insights or staying within budget. That tension creates a real bottleneck on the path to becoming AI-ready.

Our goal with the Tricentis blog is to distill insights that help QA professionals navigate the massive, AI-driven transformation happening across the software delivery landscape. To that end, I reached out to experts across Tricentis, from product and services to marketing and strategy, to hear what they’re really thinking about AI in QA right now. This group brings decades of experience building testing products, guiding enterprise transformations, and shaping how organizations adopt AI.

70% of teams are already integrating generative AI tools into their daily workflows, according to our 2025 State of Game Technology Report. Now more than ever, teams are looking to connect their AI tools to the services and applications they rely on to get work done. To address this issue, the industry has begun to standardize using the Model Context Protocol (MCP) to connect their existing tools and LLMs like Claude, GPT, and Gemini.

Modern AI applications are no longer just about sending prompts to an LLM and returning text. As soon as AI systems need to interact with real business data, internal APIs, or operational workflows, the problem becomes one of orchestration, security, and control. The challenge is to build secure AI agents without embedding fragile logic or exposing sensitive systems directly to a model. This is where a layered architecture using Volcano SDK, DataKit, and Kong MCP Proxy becomes compelling.

AI transforms CI/CD testing from reactive bug detection into proactive quality assurance that accelerates release cycles while improving software reliability. Start embedding AI into your testing workflows now because teams that wait will struggle to match the velocity of competitors who already have. Continuous integration and continuous deployment pipelines have become the backbone of modern software delivery.

Orchestrating distributed AI workloads Distributed (multi-node) training has become a requirement rather than an optimization for many modern AI workloads. As model sizes grow, datasets expand, and training timelines tighten, teams increasingly rely on multiple machines, often with multiple GPUs each, to complete training efficiently.

Software testing was once a slow and repetitive process that developers accepted as unavoidable, often consuming significant time without delivering proportional value. Traditional manual testing struggled to scale with growing application complexity and rapid release cycles. In 2026, test generating tools have reshaped this landscape by introducing automated test generation, AI-driven logic, and intelligent coverage strategies.

A customer recently shared their debugging workflow with me. When an error shows up in Honeybadger, they import it to Linear, manually add context about where to look in the codebase, then assign GitHub Copilot to investigate. It works, but they asked a good question: could Copilot just access Honeybadger directly? The answer is yes—and it's easier than I expected.

Have you ever asked an AI assistant to generate code for a framework it doesn't quite understand? Maybe it produces something that looks right, but the syntax is slightly off, or it uses deprecated patterns. The AI is working hard, but it lacks the specific context it needs to truly help you. The Model Context Protocol (MCP) was designed to bridge this knowledge gap by giving AI assistants access to domain-specific knowledge and capabilities they don't have built in.