Build the right thing.
Question assumptions, understand users, and decide whether to build, buy, connect, or simplify before adding more software.
Have a software idea or need technical leadership? Reach Kevin directly.
Getting an idea onto a screen is more accessible than ever. Turning it into software your business can depend on takes context, engineering, and someone accountable for the result.
WHY INVOLVE AN ENGINEER?
Building with AI yourself can be a great way to explore an idea, learn what you need, and test an early workflow. If you have already started, bring what you have. The next step is understanding what is ready, what needs attention, and what is worth building further.
Real users introduce questions a demo can leave unanswered. Who can access each record? What happens when an integration fails? Can data be recovered? Will the system remain understandable as it changes? An experienced engineer knows to ask those questions and can work through the answers with you.
WHAT THE HUMAN ADDS
Question assumptions, understand users, and decide whether to build, buy, connect, or simplify before adding more software.
Review architecture and data access, test important behavior, and handle failure paths alongside the happy path.
Plan for deployment, monitoring, recovery, maintenance, and the next person who will need to understand the system.
HOW THE WORK HAPPENS
Understand the current process, constraints, and what improvement would be valuable. Define how we will assess the result.
Make the scope and tradeoffs clear. A focused review, a prototype, an integration, or a roadmap may be the best starting point.
Use AI where it helps exploration and implementation. Review generated work, verify key behavior, and make deliberate decisions about sensitive information and tool access.
Check the result against the agreed goal, explain limitations, and establish ownership and the next steps for support or continued development.
LESS WORK FOR PEOPLE
The aim is to remove unnecessary tasks, decisions, and handoffs. Use existing systems and channels wherever practical. Keep people involved for approvals, judgment, and exceptions, with clear controls and a way to review what happened.
For a customer-facing product, that means an experience that helps people accomplish something. For an internal system, it can mean fewer logins, fewer status checks, and a useful notification only when someone needs to act.
STARTED WITH AI? STILL AT THE IDEA STAGE?