Identify the Problem
Define what is not working, what outcome matters, and whether the problem is worth solving.
Process
How I Work
I do not start by asking, “What can AI do?” I start by asking, “What problem are we trying to solve?”
Understand the Workflow
Map the current process, users, data, exceptions, dependencies, bottlenecks, and constraints.
Design the Solution
Choose the simplest useful architecture. Not every problem needs new software.
Build With AI
Use AI to accelerate research, requirements, interface design, coding, content, analysis, debugging, and documentation.
Connect the Systems
Integrate the databases, APIs, communication tools, forms, external services, and existing software the workflow depends on.
Test and Improve
Use realistic workflows, find edge cases, fix failures, simplify the experience, and improve based on actual use.
The principle
Technology follows the problem
The goal is a useful working system, not a complicated stack.
If a process change solves the problem, do that. If an existing tool can be configured, use it. If systems need to connect, integrate them. If the workflow needs custom software, build it. AI makes more of those options practical and faster to test.
Talk About a Project