Replacing manual outsourced QA with AI agents

Company Overview

• Company: ANY.RUN
• Industry: Cybersecurity / Threat intelligence
• Business Type: Interactive malware analysis sandbox platform
• Size: 50–100 employees
• Headquarters: Global

Key Use Cases

• Automated marketing QA (landing pages and complex forms)
• Pre-production environment verification with AI (via secure tunnels)
• Shift-left pull request (PR) testing
• Testing artifact & result generation

The Challenge: Hyper-growth and the limits of outsourced manual QA

ANY.RUN is a fast-growing cybersecurity company that provides an interactive malware analysis sandbox platform. Over the last year and a half, the organization experienced massive growth, tripling in size.

To keep up with the overwhelming amount of manual verification required for new releases, the company relied heavily on hiring offshore manual QA contractors from regions like India, Bangladesh, and Eastern Europe. However, this approach presented significant scaling issues. Relying on manual labor was highly prone to human error, lacked predictability, and failed to provide management with clear, transparent artifacts showing what was actually tested.

Radik Zagirov, who is driving the QA transformation at ANY.RUN, needed a way to replace this manual dependency with a highly predictable, automated solution – without the massive hidden maintenance tax of writing thousands of traditional scripts.

The Solution: Quick wins in marketing and shift-left PR testing

ANY.RUN decided to partner with QA.tech to systematically replace their manual QA contractors with AI agents, starting with a strategic "quick win" approach.

Targeting marketing QA first:

‍Because ANY.RUN is a cybersecurity company, testing their core malware sandbox and drivers involves complex infosec hurdles. To bypass these initial blockers and prove ROI immediately, Radik deployed QA.tech agents to automate the marketing department's testing – specifically focusing on landing pages, forms, and bot-protection flows in non-production environments.

Secure tunnels for pre-prod:

ANY.RUN leverages QA.tech's tunnel functionality to securely test deployments on their internal, non-production staging environments before features are released.

Integrating into developer PRs:

‍Looking beyond marketing, ANY.RUN is actively building a business case to integrate QA.tech directly into their developer pipelines. By having AI agents automatically test PRs, developers receive immediate feedback, further reducing the need for manual QA intervention.

The Results: High ROI and complete transparency

By introducing QA.tech, ANY.RUN is transitioning its existing in-house QA leads from manual clickers into AI supervisors who manage automated workflows.

Management immediately saw the value in the platform's visual approach. As Radik Zagirov noted:
‍"It's always a question with the quality assurance, where's the artifact, where are the results of your work... everyone loved the fact that it's very transparent, you can see the results, you can share them".

Ultimately, ANY.RUN is proving that AI-driven testing is not only more reliable than outsourced manual labor, but also significantly more cost-effective. Radik estimates the potential return on investment for replacing outsourced contractors with QA.tech agents is massive.

ANY.RUN Highlights

• Replaces outsourced manual QA with AI agents to support 3x organizational growth.
• Uses secure tunnels to run automated tests on pre-production environments.
• Provides management with clear, shareable testing artifacts.
• Delivers an estimated ROI of up to 900% compared to manual testing contractors.