Enterprise Software
13 posts on this topic — practical guidance from Shawn Livermore on fractional CTO, AI, and technology leadership.
The AI Automation Architecture Mistake Most Companies Make
Most AI automation projects fail not from bad tools but from bad sequencing. Here is why architecture must come before tool selection — and what happens when it does not.
Read post →Why the Enterprise Software Delivery Model Changed Before the Architecture Did
AI coding tools changed how fast enterprise software gets written. The harder adjustment is to the architecture, the governance, and the leadership model that sustains it.
Read post →AI Now Outperforms Junior Accountants. The Gap to Enterprise Deployment Is the Real Story.
Frontier AI models are faster and more accurate than junior accountants on well-defined accounting tasks. What Ethan Mollick's research means for organizations trying to make that result reliable inside their actual workflows.
Read post →Building Enterprise Software in the AI Era Is a Different Job
Andrej Karpathy's Sequoia AI Ascent 2026 talk made the point clearly: LLMs create new capabilities, they don't just accelerate old ones. Enterprise software creation is no exception.
Read post →Software Is Getting Cheaper to Build. That Doesn't Make the Build Decision Simpler.
Jason Lemkin replaced a $4,000/year tool with something built in an hour on Replit. That's real. The question business owners need to answer is which of their software actually falls into that category.
Read post →The Harness Is the Strategy: Enterprise Software in the AI Era
The 2026 AI engineering consensus is clear: the model is not the bottleneck. The harness — evaluation, context management, orchestration, permissions — is what separates reliable software from expensive experiments.
Read post →Enterprise Software Creation in the AI Era: What Actually Changes
Karpathy's LLM Knowledge Base concept — using AI to compile and maintain institutional knowledge rather than just generate code — is the most significant architectural signal for enterprise software teams in 2026. Most organizations won't act on it until they're behind.
Read post →Your Vibe Coding Starter Template Is an Architecture Decision
When you pick a starter template for a vibe coding project, you are making the first architectural commitment of the entire codebase. Everything the AI generates afterward works within those assumptions.
Read post →The Strategic Shifts in Enterprise Software Creation That AI Is Actually Forcing
AI tools produce more code, faster. The harder question is what that means for architecture governance, build-vs-buy decisions, and team structure — the decisions that were already difficult.
Read post →The Enterprise Software Specification Is Broken. AI Is Making That Visible.
AI coding tools fill specification gaps with defaults from their training data. In enterprise software, those defaults are rarely correct. The specification problem that always existed is now impossible to ignore.
Read post →Build vs. Buy Software in the AI Era: When the Economics Shift
AI has moved the cost curve for custom software development. That changes which software is worth building. Here is how to re-evaluate the build vs. buy decision when your development cost assumptions no longer hold.
Read post →What Anthropic's Parallel Subagents Mean for Enterprise Modernization Programs
Claude Opus 4.8's Dynamic Workflows can run hundreds of parallel subagents across a codebase. The bottleneck in enterprise modernization has never been coding speed — it's architecture and sequencing. Understanding the difference changes how you use the capability.
Read post →Before Your Team Vibe Codes, Define the Code Footprint
AI coding tools generate code faster than teams can review it. Quality is set before the first prompt, by your starter template, context file, and defined target.
Read post →