Useful AI needs more than a prompt. I build products and workflows that connect it to your business.
SEE HOW ↓khaled@os %direct: take everything. plan it. build it.
Prompt AI, get answers. Architect agents, get systems — connected to the work that needs doing.
6COMPLETED BUILDS
FOUR FEATURED CASE STUDIES · TWO SUPPORTING WORKFLOWS
PROJECT DESCRIPTIONS AVAILABLE · DEMO MEDIA COMING NEXT
AI products, customer-facing assistants and business automation. Built work, explained through the problem it solves. Demo recordings and screenshots are coming next.
NO.01Prospect information is scattered across sources and arrives in inconsistent formats. A useful outreach dataset needs relevant context and a consistent structure.
I built an n8n lead collection and enrichment workflow that aggregates prospect information, normalizes fields and filters records against an ideal customer profile.
Routine inbox processing competes with more useful work: identifying a message's purpose, applying labels and preparing a response.
I built email-management automations using n8n for classification, labeling, AI-assisted drafts and configured replies. These are supporting workflows, separate from the full outbound case study.
A submitted form is only the start. Follow a lead through qualification, response and routing. This is a workflow explanation, not a replay of customer data.
A lead submits the form.
Collect the business information and qualifying answers needed to evaluate fit.
WORKFLOW OVERVIEW · NO CUSTOMER DATA
INBOUND LEAD QUALIFICATION · ILLUSTRATIVE WORKFLOW
The interface is one part of the work. The connections behind it make the system useful.
ReFindLink connects browser capture, content extraction and AI processing to a searchable library.
The commerce assistant connects product questions and visual lookup to catalog information and order records.
Inbound qualification routes form submissions. Outbound outreach connects prospect research to email and reply tracking. They are separate completed systems.
Explore the main components of each completed system. Different business needs, different implementations.
Products, AI assistants and workflow automation. Choose a project to see its components.
Ask about the completed projects and what they do. Answers come from the project notes, without a live AI model.
Five illustrative engineering scenarios. Explore the tradeoffs and make the call. These are exercises, not client incident reports.
Map demonstration only. These routes are not a record of my travel.
Bring the business problem. We can scope the product, assistant or workflow around what your team needs.
Web applications, browser extensions and dashboards with AI built into the workflow.
Assistants that connect customer questions to product information and operational actions.
Lead qualification, prospect research, outreach and email processing connected to your existing tools.
Start with one business problem. Build the system around it.
