Agentic AI transforms QA by automating testing tasks with large language models, reducing reliance on coding expertise but facing challenges like skill gaps and maintenance. QA professionals must adapt through AI training, balancing automation with human oversight to ensure accurate testing and career growth.

How To Test This?

interview episodes where Mamadou N’diaye talks with with software testing experts

• https://podcasters.spotify.com/pod/show/spidey1944
• https://www.youtube.com/@HowToTestThis
• https://www.linkedin.com/in/mamadou-ndiaye-consultant/

Episode Details

• Show Notes: https://podcasters.spotify.com/pod/show/spidey1944/episodes/How-to-Test-With-Agentic-AI-automation-Tool---Geosley-Andrades-e3h52pi
• Published: 2026-03-29T21:30:17Z
• Duration: 00:50:33
• Author: Mamadou N’diaye

Overview

The podcast explores the impact of agentic automation AI on quality assurance (QA) teams, emphasizing its role in transforming traditional testing practices. Agentic AI leverages large language models (LLMs) to automate tasks like test case generation, synthetic data creation, and code writing, reducing reliance on deep coding expertise. Key features include autonomous discovery, where AI analyzes application interfaces to identify issues, and self-healing capabilities that adapt to evolving applications, minimizing maintenance burdens. However, challenges persist in the testing industry, such as skill gaps in coding and AI, maintenance complexities with conventional frameworks, and scalability issues in manual processes. To address these, experts recommend starting with small AI implementations, collaborating with domain experts, and investing in training to bridge knowledge gaps.

The discussion highlights the need for strategic AI adoption, focusing on tools that support end-to-end testing across web, APIs, and enterprise systems (e.g., Salesforce, SAP) while prioritizing design principles and contextual intelligence to avoid hallucinations. Misconceptions about AI include over-reliance on automation without proper validation or context, which can lead to flawed outcomes. Effective use requires balancing AI-generated outputs with human oversight, ensuring alignment with business logic and testing rigor. Enterprise applications face unique challenges, such as frequent UI updates disrupting automation scripts, which agentic platforms like ExcelQ address through live asset updates and pre-built accelerators. The conversation also underscores the importance of metrics like bug quality, automation efficiency, and ROI evaluation to guide AI tool selection and implementation.

Finally, the podcast stresses the evolving role of QA professionals in an AI-driven era. While agentic systems accelerate testing, human expertise remains critical for prompt engineering, validating AI outputs, and understanding design principles. Career advice emphasizes continuous learning, adaptability, and mastering AI-specific concepts (e.g., RAG pipelines, vector databases). Testers are encouraged to embrace AI as a tool to enhance efficiency rather than replace their roles, focusing on contextual awareness and fast-paced testing to align with modern development cycles. The shift from traditional methods to AI-integrated testing demands a balance between automation and human judgment, ensuring robust, scalable QA frameworks.

What If

• What if you pilot an agentic AI tool for test case generation in your next project?

  • Concrete move: Use a platform like ExcelQ to automatically generate test cases for a complex enterprise application (e.g., Salesforce or SAP).
  • Why now: Your current manual testing is slow and error-prone, and agentic AI can adapt to dynamic interfaces, reducing maintenance costs by 50% (per ExcelQs live asset updates).
  • Expected upside: Faster setup, reduced flakiness in test scripts, and scalable automation for future features.

• What if you create a personal resource repository for agentic AI prompts and templates?

  • Concrete move: Build a “cheat sheet” of pre-defined prompts for tasks like synthetic data generation, self-healing scripts, or API test case creation.
  • Why now: The text emphasizes the need for prompt engineering and structured templates to avoid hallucinations or inefficiencies when using AI tools.
  • Expected upside: Streamline repetitive tasks, reduce trial-and-error time, and ensure consistency in AI outputs across your workflows.

• What if you run a proof-of-concept (POC) using no-code agentic automation for a high-maintenance task?

  • Concrete move: Apply a no-code solution like XLQ to automate a complex enterprise workflow (e.g., login + data validation in an Oracle app).
  • Why now: The discussion highlights that no-code tools can bridge skill gaps and handle frequent UI updates, which your current framework struggles with.
  • Expected upside: Validate ROI with minimal resource investment, demonstrate value to stakeholders, and identify pain points for scaling later.

Takeaway

• Start with small proof-of-concept (POC) projects to evaluate agentic AI tools, focusing on specific use cases like test case generation or synthetic data creation before scaling adoption.
• Invest in training for AI/ML basics and agentic workflows to upskill your QA team, ensuring they can effectively collaborate with AI tools and validate outputs.
• Prioritize AI platforms that support end-to-end automation across web, APIs, databases, and desktop apps, avoiding fragmented tools that require manual integration.
• Evaluate AI tools using real-world metrics such as bug quality (not just quantity), automation time, and defect slippage to ensure tangible ROI and efficiency gains.
• Implement precise prompting strategies and validate AI outputs with contextual data to avoid hallucinations, ensuring synthetic test cases, code, or analysis align with business logic.