- Able Craft & Code
- Posts
- How to manage your SDLC when AI changes everything
How to manage your SDLC when AI changes everything
Turns out the hardest part of AI adoption isn't technical—it's human
In this issue:
CRAFT: Learn what happened when AI nearly broke our SDLC
CODE: Find out how we used AI to refactor a complex legacy code
NEWS: OpenAI steals Apple's star designer, Walmart is training for robot shoppers, and Claude unveils models that can code for hours
LEARN: Sign up for our webinar to learn about CTO best practices for AI implementation
CRAFT NOTES
What we’re thinking about
Hi tech leaders,
Our team is pretty fearless when it comes to experimenting with new technology. In fact, it's the thing that drives us. But a year ago, we hit an unexpected roadblock as we were refining our SDLC to incorporate AI tools and methodologies.
Frankly, it was a humbling moment. We tackle big technical challenges all the time. And we'd been planning for AI use cases long before we were able to get our hands on the tools. We hit the ground running.
But this roadblock wasn't technical—it was human.
Suddenly, everyone's roles started shifting. Engineers were generating documentation that product managers used to own. Product managers were prototyping features that used to require full dev cycles. QA was happening in places we'd never seen it before. Everyone was stepping into roles that didn't exist six months earlier.
We knew AI would change how people worked. What we didn’t anticipate was how much it would change how they worked together.
We needed to redefine our development process. This led to some uncomfortable conversations about roles, responsibilities, and relevance. Ultimately, we fell back on our tried and true approach: We created structured experiments to see what worked, what didn't, and where we could make the biggest gains.
Then something interesting happened. The more we implemented AI internally and with our partners, the more we found that the "magic" isn't in AI tools themselves—it's in how engineering roles are reshaped around them.
As our Director of Engineering Pradeepa Dhanasekar recently noted on our blog, we're witnessing a shift from "syntax mastery to systems thinking."
This shift has been particularly relevant for our midmarket partners. While enterprise companies often have dedicated AI teams and startups can pivot quickly, midmarket companies face a unique challenge: they need the efficiency gains that AI provides, but they can't afford to disrupt existing workflows or lose institutional knowledge during the transition.
In this edition of Able Craft & Code, we're sharing the specific experiments and frameworks that helped us turn our unexpected roadblock into an innovation accelerator—insights that have proven especially valuable for growing companies ready to implement AI without breaking what already works.
–Andy McKinney, CEO
CODE LAB
What we’re discovering
"Tools like Cursor and Copilot are helping engineers write code faster. They are debugging more efficiently and automating a lot of tasks. In one of our recent projects, our team was able to refactor a complex legacy code base that has been problematic for months. So what previously would have taken around three sprints of work is now just completed in one, with AI handling the repetitive transformation patterns while we focus on edge cases. So I believe this acceleration is giving engineers more time for creative problem solving."
Why this matters for your team: Your developers are likely spending up to 40% of their time on repetitive code transformations instead of solving core business problems. By strategically applying AI tools to handle pattern-based refactoring tasks, you can dramatically compress development cycles while maintaining quality and freeing your best talent to focus on innovation rather than maintenance.
AI NEWS
What we’re paying attention to
OpenAI Drops $6.5B to Steal Apple's Star Designer
OpenAI just completed its largest acquisition ever, paying $6.5 billion in stock to acquire Jony Ive's AI device startup "io" and secure the legendary iPhone designer to lead hardware development for next-generation AI devices. This strategic move signals OpenAI's ambition to own the hardware platform rather than depend on iOS and Android distribution, potentially reshaping how users interact with AI and creating new competitive dynamics in your market by 2026. (Source: Bloomberg.)
Walmart Preps for the Rise of AI Shopping Agents
Walmart is building its own shopping agents while preparing for third-party AI bots like OpenAI's Operator to autonomously browse and purchase products on behalf of consumers. Since these agents may be less attracted to emotional visuals and instead prioritize factors like search rankings and pricing, retailers will need to adapt their product descriptions, pricing strategies, and site optimization for algorithmic rather than human decision-making. (Source: Wall Street Journal.)
Anthropic Says New Claude 4 Can Think for Hours
Anthropic just released Claude Opus 4 and Sonnet 4, with Opus 4 leading on coding benchmarks like SWE-bench (72.5%) and demonstrating the ability to work continuously for several hours on complex tasks. The models introduce "extended thinking with tool use," letting Claude alternate between deep reasoning and web searches, while new memory capabilities allow it to maintain context and build knowledge over time—potentially transforming how your development teams approach long-term projects. (Source: Anthropic.)
LEARNING OPPORTUNITIES
Webinar - The human-centered AI roadmap: frameworks technology leaders need for successful AI adoption

The most successful AI integrations aren't technical projects—they're leadership challenges. Discover the frameworks that help CTOs navigate role evolution, maintain team cohesion, and accelerate AI adoption. Join Pradeepa Dhanasekar, Director of Engineering, Jose Rodriguez, AI Solutions Lead, and Fede Garcia, Senior Software Engineer as they share real experiments and practical strategies for building teams where AI amplifies human capability.
June 24
11-11:45 am PST
SKIP THE ENDLESS EXPLORATION, START BUILDING
Done with endless AI meetings? Ready for practical implementation?