RoleModel Software

Meet this year's Craftsmanship Academy cohort. They're just getting started, and the real learning happens in the work itself. Stay tuned over the next few weeks as they dig into real-world projects and hands-on learning.

Wishing everyone a happy 4th of July as we celebrate America's 250th. We are proud to call this country home!

Custom software isn't a cost to justify. It's a resource to deploy — like hiring the right person for the right job. The real question isn't "can we afford it?" It's "are we putting our resources to work on something that matters?" Check out our blog post where we talked to Jordan Raynor about the stewardship mindset behind smart software decisions. https://lnkd.in/e8gTJx9Z

Are you treating AI like a slot machine? You ask a question, get a response and move on. There's a better way! Download our free guide today! https://lnkd.in/eJbqJyx9

How do you weigh the output you get from an AI tool? Question everything? Blind trust? I'm finding that two axes matter when calibrating how you evaluate AI output. The first is what's at stake. Something for your own thinking needs a quick sanity check. Something shared with your team needs a structural review. A production deliverable needs source verification on every claim. The second is what type of work the AI produced. Qualitative work (writing, structure, framing) is where the model is generally reliable. Quantitative work (numbers, data, percentages) is where it can struggle. Factual work (claims, citations, current data) is where it fabricates or goes stale. Don't over-verify low-stakes work. That's waste. Don't under-verify high-stakes work. That's risk. Calibrate to your situation.

The Academy is a starting point, not a finish line. Graduates from our 2024 and 2025 cohorts share where their craftsmanship journey has taken them since completing the program. The next cohort begins this summer. Check back to follow their journey. #customsoftware #businesssoftwaredevelopment #businesssolutions #businesstools 

How do you know when it's the right time to invest in software? Not too early. Not too late. Caleb Woods talked through exactly that with Jordan Raynor, bestselling author of Redeeming Your Time. If you want more thinking like this on faith and work, Jordan's weekly devotional is worth your time: https://lnkd.in/eaKUAprs

Most teams are focused on producing work faster with AI. The real expertise amplification is capturing what is learned in each session and carrying the right learning forward. Rick Reilly at Test Double wrote a piece that nails this: "Red-Green-Refactor Your Context" (https://lnkd.in/ekB2YCZV). The premise is simple. Treat the knowledge you gain during an AI coding session with the same discipline you bring to TDD. Capture the problem before solving it. Document decisions as you go. Then route what you learned to where it belongs, whether that's a README, a script, or a project convention. I've always seen TDD as being about two things: confidence and documentation. I wrote about why in "The RoleModel Way of Testing" (https://lnkd.in/ee_S7Wcj). Self-documenting code. Executable business use cases. Tight feedback loops that tell you when something breaks. AI makes it possible to expand what kind of documentation is created, but it still has to be managed well to keep the feedback loop tight. One habit I've been experimenting with is ending every AI session by asking it to update the project's instructions with what it learned. It takes seconds. And over time, each session is more tailored to my workflow and use case. James Clear said it well in Atomic Habits: "Small habits don't add up. They compound." This is a critical piece of AI leadership that is just starting to get attention. It isn't just about prompting better, it is about building systems to make every AI session better for each member of the team. How are you baking what you learn in each AI session into something that supports team collaboration?