Keyhole Software

Most developers think of GitHub Copilot as a coding assistant. But one of its most valuable features might actually be something else: pull request reviews. In his latest article, Jonny Hackett shares how GitHub Copilot has become another reviewer in his development workflow, providing fast, consistent feedback before a teammate ever opens a PR. He covers: ✅ Where Copilot PR reviews provide the most value ✅ Common issues it can identify ✅ Limitations to be aware of ✅ Best practices for development teams ✅ Why human reviewers remain essential One of my favorite takeaways is that Copilot isn't there to replace developer review. It's there to surface potential issues, encourage better discussions, and allow reviewers to spend more time on architecture, maintainability, and business context. If your team is evaluating how to incorporate AI into your software development workflow, this is a practical, experience-based read. 📖 Read the full article: https://lnkd.in/g_3-jQ7d #GitHubCopilot #CodeReview #SoftwareDevelopment #DeveloperTools #EnterpriseAI

If you haven't paid attention to Java recently, you might be surprised by how much has changed. Virtual threads, pattern matching, records, cloud-native tooling, and continued innovation across the Spring ecosystem are reshaping how modern Java applications are built and maintained. At the same time, organizations are balancing modernization efforts with the reality that many of their most important systems still run on Java. In our latest article, we explore the Java trends shaping 2026 and what they mean for developers, architects, and technology leaders. Read the article: https://lnkd.in/gwU7bMYZ

Ever lose your place after refreshing a log viewer, search result, or admin screen? As APIs scale, many are moving from offset pagination to cursor-based pagination for performance reasons, but that shift comes with user experience tradeoffs. In her latest article, Rachel Walker explores: ✔️ Cursor-based pagination vs offset pagination ✔️ Why modern APIs are making the shift ✔️ Keyset vs session-based cursors ✔️ Techniques for preserving navigation and shareable links ✔️ An approach to restoring familiar navigation patterns with anchor-based navigation Read the article: https://lnkd.in/gi_rgK7K If you build APIs, frontend applications, or large-scale systems, this is a great look at the tradeoffs between scalability and usability. Thanks to Rachel for a thoughtful deep dive into a topic that affects both developers and users. #SoftwareArchitecture #APIDesign #GraphQL #WebDevelopment #FrontendDevelopment #UserExperience #PerformanceOptimization

Everyone wants to talk about AI. Fewer organizations want to talk about the systems standing in the way of it. Many enterprise applications were built long before cloud-native architectures, real-time integrations, modern security expectations, or AI-powered workflows became part of the conversation. Yet these same systems continue to run critical business operations every day. As organizations look to take advantage of AI, improve security, increase agility, and reduce operational risk, those legacy systems are receiving renewed attention. We took a look at the latest market data, adoption trends, and growth projections shaping modernization efforts in 2026. Read more: https://lnkd.in/gPd9Y58f #LegacyModernization #ApplicationModernization #AI #CloudMigration #DigitalTransformation #EnterpriseArchitecture

What does AI-assisted legacy modernization actually look like in a real client environment? Evan Sanning just shared a behind-the-scenes look at a recent Keyhole project rewriting a legacy Delphi application into a .NET service architecture for a pharmacy client, completing the project 40% faster than the original delivery target. What I appreciate most about this write-up is that it skips the usual "AI hype" and focuses on what using LLM tools inside a real engineering workflow actually requires. The biggest takeaway: the project succeeded because of upfront architecture, planning, and engineering judgment. The LLM tools accelerated implementation once the structure and workflows were already in place. He also dives into: - Context window management - Hallucinated implementations - Migration risks hidden in legacy code - Workflow patterns using AGENT.md, MEMORY.md, and ARCHITECTURE.md Excellent work by Evan documenting this process so thoughtfully. Read the full article here: https://lnkd.in/gEj4EAUj #LegacyModernization #SoftwareModernization #DotNet #Delphi #AI #LLMs #GitHubCopilot #SoftwareArchitecture #EnterpriseSoftware

Keyhole Software was recognized in the Clutch Top 15 for 2026 across software development and application modernization. That is exciting, but what matters more is what those rankings actually reflect. The companies that consistently deliver in enterprise environments are not defined by the tools they use. They are defined by how they approach: • Architecture that holds up over time • Modernizing complex systems without disrupting the business • Applying AI to accelerate delivery without introducing risk • Integration into complex, real-world systems We took a closer look at what is behind those patterns and what it means for organizations evaluating software development partners in 2026. If you’re thinking about modernization or where AI fits into your delivery approach, this is worth a read: https://lnkd.in/gfw5jSrp #SoftwareDevelopment #ApplicationModernization #AI #EnterpriseTechnology

Lou Mauget has been writing code since 1961, starting in machine language on the ILLIAC I. Today, he’s building in React and TypeScript and exploring how AI fits into modern development workflows. What stands out is not just the length of his experience, but that he continues to stay at the forefront of each new wave of technology, not from a place of hype, but from a genuine curiosity and experience that has evolved over six decades. It’s not about whether AI can help. It’s about how to use it practically within a real development workflow - in this case, generating and refining unit tests in a React and TypeScript application using Vitest, React Testing Library, and GPT-based tools. If you’re thinking about where AI fits into your development process, this workflow is well worth a read: https://lnkd.in/gHphKXtc #AI #UnitTesting #TestAutomation #ReactJS #TypeScript #DeveloperTools #AIDevelopment

Most teams say they’re "𝒅𝒐𝒊𝒏𝒈 𝑨𝑰" in software development right now. What that usually means: they have Copilot licenses. What it doesn’t mean: they’re getting consistent, repeatable results. Over the past few months, we’ve been comparing notes across our engineering teams and client work. A few patterns keep showing up: - We’re often brought in to fix AI-generated apps that looked complete early, but don't hold up under real usage - Small inconsistencies in AI-generated code compound quickly without guardrails - Strong engineering teams get faster with AI, weak processes get exposed - Developers still review everything; AI speeds up work, not accountability - The teams seeing results are structuring work around AI, not just prompting it We pulled together what’s actually working in production: 👉 https://lnkd.in/guvjHYR4 #SoftwareDevelopment #ArtificialIntelligence #EnterpriseAI #SoftwareEngineering #AIinEngineering Curious what others are seeing, especially where AI has helped vs. where it’s created more cleanup.

If you’re building with FastAPI and PostgreSQL, managing schema changes gets tricky fast. The #KeyholeTeam's James Fielder put together a hands-on guide to using Alembic with FastAPI and SQLModel, including: • Migration setup and workflow • Rollbacks and data seeding • Docker-related gotchas that tend to trip people up It’s a straightforward, real-world example that shows how to manage schema changes as your application evolves. Full walkthrough here: https://lnkd.in/gMw3kWwv #FastAPI #Python #PostgreSQL #Docker #SoftwareDevelopment #APIDevelopment

We spend a lot of time helping teams think through modernization decisions. Not just how to modernize, but what approach actually makes sense for their systems and constraints. The key decisions are what to change, what to keep, and how far to go. Not everything needs to be rebuilt, but not everything should be carried forward either. Modernization is a series of tradeoffs across risk, speed, and long-term outcomes. We put together a guide to help teams work through those decisions more intentionally: • Evaluating system constraints and goals • Choosing the right modernization path • Applying AI in a structured, low-risk way • Building a clear path forward If you’re thinking about modernization this year, this is a good place to start: 👉 https://lnkd.in/dMDWUHXw #SoftwareModernization #DigitalTransformation #CloudModernization #EnterpriseArchitecture