I approach AI the same way I approach other enterprise technologies: start with the business problem, understand the information and controls required, select an architecture appropriate to the risk and value, and validate assumptions before adding unnecessary complexity.
My AI experience includes enterprise architecture and governance as well as hands-on personal experimentation.
In my current Enterprise Architecture work, I contribute to architecture and governance for AI-enabled capabilities and collaborate with teams working to connect generative AI with enterprise knowledge and business context.
A recurring challenge with enterprise AI is that the model rarely has enough context by itself.
Understanding a business question may require company-specific definitions, calculations, applications, dashboards, security context, and authoritative data.
That makes knowledge grounding, architecture, governance, and integration at least as important as the choice of language model.
Ask Jerry is a personal AI project I designed and deployed to allow recruiters, hiring managers, executives, and professional contacts to ask natural-language questions about my professional experience.
It is intentionally not a general-purpose chatbot.
The goal was to build a narrowly scoped system capable of answering questions from verified career information while reducing hallucinations, unsupported conclusions, privacy problems, and unnecessary complexity.
You can try it by clicking "Ask Jerry AI" on the menu.
ShowNow is a personal software project I created to explore how generative AI can change the way applications are designed and built.
The application addresses a real-world coordination problem in residential real estate: scheduling home showings among buyers, agents, sellers, and other authorized approvers. ShowNow brings showing requests, approvals, alternate-time suggestions, notifications, showing history, and feedback into a structured workflow.
One of my primary goals with ShowNow was to gain hands-on experience with AI as part of the software-development lifecycle.
I used ChatGPT to help refine requirements, reason through architecture, identify edge cases, troubleshoot problems, develop test scenarios, and create detailed implementation instructions.
I then used Cursor as a repository-aware AI development environment to inspect the existing application, generate and refactor code, execute builds and tests, diagnose failures, and iteratively implement changes.
My role remained that of product owner, architect, technical decision-maker, tester, and reviewer. AI accelerated implementation, but decisions around domain design, security, integrations, privacy, architecture, and whether the resulting software actually met the requirement remained human responsibilities.
ShowNow is built primarily on ASP.NET Core/.NET, Entity Framework Core, SQL, and ASP.NET Core Identity.
The application also gave me direct experience integrating and designing around several cloud and SaaS services:
Resend for transactional email, Firebase Cloud Messaging for push notifications, Firebase Analytics, Twilio for SMS, and Cloudflare Tunnel and DNS for securely exposing the application without traditional inbound port forwarding.
Rather than tightly coupling the application to individual vendors, I designed provider abstractions around areas such as notifications, property/listing data, photo storage, and mapping/routing. That makes individual services replaceable without requiring the core business logic to be rewritten.
ShowNow reinforced something I believe will become increasingly important for technology leaders: AI can dramatically reduce the distance between an architectural idea and working software, but it does not eliminate the need for architecture or exeperience developers.
In fact, AI-assisted development makes clear requirements, security boundaries, integration patterns, understanding engineering standards, testing, and human code review even more important.
A vague instruction can produce code very quickly. A well-designed architecture combined with detailed constraints and acceptance criteria can produce useful software very quickly.
For me, ShowNow has been an opportunity to move beyond evaluating AI conceptually and gain direct experience incorporating AI into the engineering process—from product idea and architecture through implementation, integration, debugging, testing, and deployment.
This is a beta but you can visit ShowNow here.