Operations Architect
1 Week
AI Infrastructure
n8n, WordPress, Webhook, AI Agent, API, etc
A portfolio site has one primary goal: turning visitor interest into a conversation.
Most portfolios fail at the exact moment that matters most. A visitor reads a case study, becomes interested, and then encounters a contact form that requires them to stop exploring and start filling out fields. The momentum disappears before the message is ever sent.
The problem on my own WordPress portfolio was not discoverability. The content was there. The case studies were detailed. The challenge was creating a bridge between curiosity and contact.
The site successfully attracted visitors, but there was no mechanism to capture intent at the moment interest turned into consideration. Visitors who had questions about my experience, projects, or capabilities needed to navigate away from what they were reading to find answers. Every additional click introduced friction and increased the likelihood that the conversation would never happen.
The objective was to create an AI-powered assistant that could answer questions in real time, provide context about my work, and guide interested visitors toward either sending a message or booking a call without disrupting their browsing experience.
The chatbot’s primary knowledge source was my published portfolio content, including my About page and individual project case studies. These sources provided detailed context about my experience, delivery approach, and completed projects.
The system was also configured to handle questions about unpublished or NDA-protected work. Rather than attempting to retrieve confidential information, the chatbot provides high-level explanations about the types of business problems I solve, the industries I have worked in, and the capabilities I bring to client engagements.
This approach creates a useful middle ground. Visitors can still understand the breadth of my experience, even when specific project details cannot be shared publicly, while confidential client information remains protected.
The AI Agent was configured to communicate in a voice consistent with the rest of the portfolio: direct, specific, and professional without becoming overly formal.
Responses were designed to provide useful answers while encouraging meaningful next steps. The objective was not simply to answer questions but to help visitors understand whether my experience aligned with their needs.
For questions related to confidential client projects, the chatbot acknowledges NDA limitations transparently, explains the types of business problems I typically solve, and guides visitors toward a conversation if they want additional context.
The chatbot was designed to distinguish between informational conversations and contact intent.
Visitors asking about projects, services, or experience remain in a standard question-and-answer flow. Once the conversation indicates interest in working together, the workflow shifts into lead capture mode.
At that point, the system collects relevant contact information directly inside the chat experience and routes it to me through n8n without requiring the visitor to navigate to a separate contact page.
Not every visitor wants to send a message.
Some visitors prefer to schedule a conversation immediately.
Instead of presenting a static Calendly link throughout the website, the AI assistant surfaces booking options contextually when it detects that a visitor is ready to discuss a project or explore a potential engagement.
This allows the booking process to feel like a natural continuation of the conversation rather than a separate action.
The chatbot was embedded directly into the WordPress portfolio and connected to an n8n workflow through webhooks.
All processing, routing, and automation logic run on my self-hosted n8n environment deployed on Oracle Cloud Infrastructure. This ensures conversation data remains within infrastructure I control while allowing complete flexibility over workflow design and future enhancements.
A contact form can collect information, but it cannot answer questions.
I intentionally designed the chatbot to address visitor uncertainty before attempting to capture contact details. By allowing visitors to ask questions, explore projects, and understand my capabilities first, the contact process became part of the conversation rather than a separate task.
This reduced friction and created a more natural path from interest to engagement.
A common challenge for portfolio chatbots is handling questions about projects that cannot be publicly disclosed.
Rather than refusing these questions entirely, I designed the chatbot to acknowledge confidentiality constraints while still providing useful context. When visitors ask about unpublished work, the chatbot explains the nature of the restriction, describes the categories of problems and systems I typically work on, and guides the conversation toward a direct discussion if additional context is needed.
This creates a more natural experience than a simple refusal while maintaining client confidentiality.
Most websites separate discovery and contact into two disconnected experiences.
I intentionally combined them.
When a visitor demonstrates interest, the conversation naturally transitions into lead capture without requiring additional navigation, reducing the risk of losing engagement between interest and action.
Rather than relying on a third-party chatbot platform, the entire system runs through my own self-hosted automation environment.
This provides complete control over the conversational logic, lead routing process, data handling, and future enhancements while ensuring the chatbot remains fully integrated with the broader automation infrastructure supporting the portfolio.
The portfolio evolved from a static showcase into an interactive discovery experience.
Visitors can learn about my background, explore project work, ask follow-up questions, and understand how I approach problem-solving without leaving the page they are viewing. Questions that previously required navigating through multiple pages can now be answered instantly inside the conversation.
When a visitor is ready to take the next step, the chatbot routes them toward the most appropriate action, whether that is sending a message or booking a call.
The entire experience remains inside the portfolio site, reducing friction and preserving engagement throughout the visitor journey.
Most importantly, the system balances accessibility with confidentiality. Visitors can explore published projects in detail while still receiving meaningful responses about NDA-protected work. The chatbot provides context about the types of solutions I deliver without exposing confidential client information, creating a more transparent and helpful experience for prospective clients.
The result is a portfolio that not only showcases work but actively participates in converting visitor interest into meaningful business conversations.
That is exactly where I do my best work. Tell me what is frustrating you operationally, even if you cannot name it precisely yet. I will ask the right questions and tell you what I would fix first.