Wesley Mary

Lead AI Automation Architect

HighLevel • Full-time

Mar 2024 - Present • 2 yrs 6 mos

Working directly within the ecosystem of the world’s leading CRM for agencies, I focused on the transition from traditional workflows to Agentic AI models. My primary responsibility was the beta testing and deployment of AI-driven lead-handling systems for high-volume accounts. I designed and implemented the architecture for "Always-On" AI receptionists that reduced lead response times from an average of 45 minutes to under 12 seconds. By utilizing native GHL tools and external LLM nodes, I bridged the gap between basic auto-responders and complex, nuance-aware sales agents. Key Achievements: Engineered a custom "Snapshots" library that incorporated Brand-Voice AI, now used by over 500+ sub-accounts to maintain human-like authenticity in automated DMs. Reduced "Churn Rate" for SaaS-mode users by 22% by integrating automated ROI-tracking dashboards that visualized the value of AI-booked appointments in real-time. Developed a proprietary logic for "Semantic Lead Scoring" within GHL workflows, allowing the AI to prioritize high-intent leads for immediate human intervention while nurturing colder leads autonomously. Led a team of 10 developers to refine the integration between GHL’s internal AI Studio and external Search Atlas tools, resulting in a 40% increase in average indexing speed for client websites. Successfully deployed over 1,200 unique AI workflows that handled over 5 million customer interactions across SMS, Email, and Instagram without a single major system failure. My work here solidified the standard for what "Agentic AI" looks like in a CRM environment, moving the industry away from "if/then" logic toward true probabilistic reasoning and goal-oriented task execution.

Automation Architect

zapier • Self-employed

Apr 2024 - Jul 2025 • 1 yr 3 mos

Before AI went mainstream, I was a leader in the Zapier Expert community, focusing on complex "multi-step Zaps" that powered the back-end of growing SaaS companies. My work was primarily focused on "Data Orchestration" ensuring that when a lead did something in one app, five other apps updated instantly. This era was critical in developing my understanding of how data flows through a modern business, which I now use to build "Unified AI CRM Systems." I focused on the GoHighLevel integration early on, becoming one of the first experts to bridge GHL with Slack, Trello, and custom SQL databases. Key Achievements: Built and maintained over 3,000 active automation workflows for a diverse portfolio of clients, ranging from e-commerce brands to coaching businesses. Reduced manual data entry time for a major real estate agency by 40 hours per week, allowing them to scale from 10 to 50 agents without hiring additional admin staff. Won the "Automation Innovation Award" for building a system that used webhooks to trigger personalized direct mail (physical postcards) based on GHL pipeline stages. Developed a "Lead Scrubbing" automation that used third-party data enrichment tools to verify lead quality before they ever reached a sales rep, increasing the sales team's closing rate by 15%. Established a reputation for "unbreakable" workflows by implementing error-handling paths and notification systems that caught API failures before they impacted the business.

AI Implementation Specialist

ChatGPT • Part-time

Jul 2023 - Oct 2024 • 1 yr 3 mos

As an implementation specialist, I worked on the front lines of the API adoption wave, helping enterprise-level companies integrate GPT-4 and early Agentic models into their existing CRMs, specifically GoHighLevel and Salesforce. My role involved deep-level prompt engineering and "Function Calling" architecture. I was responsible for ensuring that AI agents could not only "talk" but actually "act" on behalf of the user—this involved connecting API endpoints to perform tasks like booking appointments, updating contact fields, and generating invoices based on conversational triggers. Key Achievements: Designed the "Context Injection" protocols that allowed small businesses to upload their private knowledge bases into a custom-tuned AI model within GHL, eliminating the "Generic AI Slop" problem. Optimized token usage for high-volume outreach campaigns, reducing operational costs for clients by an average of 35% while increasing the output quality of personalized messages. Authored a technical white paper on "Brand-Voice Retention in Automated Sales Funnels," which became a standard guideline for my subsequent work on Fiverr. Implemented a "Sentiment Fail-Safe" system that automatically flagged AI conversations for human review whenever a lead expressed frustration or complex technical queries, maintaining a 98.5% customer satisfaction rate across all automated interactions. Coordinated with Zapier and Make.com to build robust data bridges that allowed fragmented data from over 50 different SaaS tools to flow seamlessly into a unified GHL CRM, creating the first iteration of the "Unified AI CRM" model I now offer to clients.