Navigating the Tech Landscape: A Guide to Using the ThoughtWorks Technology Radar

Introduction

The ThoughtWorks Technology Radar is a biannual survey that captures the collective experience of the company's technologists. The 34th edition, released recently, includes 118 blips covering tools, techniques, platforms, and languages. A theme is the dominance of AI, but also a return to foundational practices. This guide will help you understand and apply the radar's insights to your own technology decisions, focusing on AI, security, and engineering discipline.

What You Need

• Access to the latest Technology Radar (free at thoughtworks.com/radar)
• Basic understanding of AI/LLMs and their role in software development
• Familiarity with software engineering practices (pair programming, testing, architecture)
• Interest in security and agent-based systems
• A journal or digital notes to record your observations and action items

How to Use the Radar: Step by Step

1. Step 1: Explore the Radar’s Structure and Blips

   Begin by browsing the radar online. It is organized into four quadrants: Techniques, Tools, Platforms, and Languages & Frameworks. Each blip has a status (Adopt, Trial, Assess, Hold) and a brief description. Take notes on blips that resonate with your current projects or interests. Pay special attention to those marked “Hold” – they may indicate technologies to avoid or revisit later.

2. Step 2: Identify AI-Related Trends

   AI dominates this edition, so filter by AI keywords. Look for blips that discuss LLM-assisted development, agentic tools, and prompt engineering. For example, the radar highlights how AI is forcing a revisit of fundamentals like clean code and zero trust architecture. Understand that AI tools can accelerate but also introduce complexity. Create a list of AI blips that are relevant to your team’s stack.

3. Step 3: Revisit Foundational Practices

   The radar emphasizes a return to core software craftsmanship: pair programming, mutation testing, DORA metrics, and accessibility. Step back from AI hype and evaluate your team’s discipline in these areas. For each practice, ask: Are we doing this well? Could it help us manage AI-generated complexity? For instance, pair programming can catch issues in AI-generated code, and mutation testing can verify test coverage.

4. Step 4: Address Security for “Permission-Hungry” Agents

   Agents require broad access to systems, but security safeguards lag. Use the radar’s security-related blips to improve your agent architecture. Consider techniques like prompt injection prevention, least-privilege design, and careful sandboxing. The radar mentions OpenClaw, Claude Cowork, and Gas Town as examples. For each agent you deploy, define explicit permissions and monitoring.

5. Step 5: Implement Harness Engineering

   Many blips in this edition focus on harness engineering – the guides and sensors that keep AI agents under control. Based on the radar meeting, this is a key theme. Identify blips that suggest specific harness components, such as monitoring dashboards, policy engines, or audit logs. Design a harness for your most critical agents. Expect the list to grow by the next radar.

6. Step 6: Document and Share Your Findings

   Create a summary of your radar exploration, focusing on 5–10 blips that have the highest impact for your organization. Share it with your team via a short presentation or internal wiki. Encourage others to add their own observations. The radar is a living document – use it as a basis for ongoing discussion, not a one-time read.

Tips for Getting the Most Out of the Radar

• Stay current: The radar is biannual – set a reminder to revisit each new edition.
• Don’t just adopt blindly: “Trial” and “Assess” mean you should experiment but be prepared to drop a technology if it doesn’t fit.
• Combine insights: Pair AI blips with security or technique blips for a holistic view.
• Engage with the community: ThoughtWorks often runs webinars on radar highlights – attend to deepen understanding.
• Use the radar as a conversation starter: Ask your team: What blips surprise us? Which ones should we test?