AI Automation Coach for CEO | Canopy Management

Date: 2026-02-18 URL: https://www.upwork.com/jobs/~022024120951217265153 Interview accept URL: https://www.upwork.com/nx/proposals/interview/uid/2024123032162249526/accept Status: SUBMITTED ✅

Client

Name: Brian (CEO)
Company: Canopy Management
Location: Austin, TX
Rating: 3.77/5 (3 reviews)
Total spent: $26K
Avg hourly paid: $72.64/hr
Size: Large company (100-1,000 people)
Hire rate: 18% (low)

Job

Rate range: $20-100/hr
Our bid: $85/hr ($76.50 after fee)
Rate increase: None
Connects: 0 (interview invitation)
Scope: Claude Code, OpenClaw, Make.com, n8n, Anthropic APIs, Supabase coaching 5-10hrs/week
Phase 2: Architect automation infrastructure for 100+ person agency

Activity

20-50 proposals, client viewed 1hr ago
Interviewing 3, invited 9, 6 unanswered
Bid range: High $906, Avg $104, Low $18

Proposal

Cover Letter

The work you described — coaching a CEO of a 100+ person agency while architecting internal automation infrastructure — is exactly what I've been building toward. I'm in.

Here's what I'll bring:

Automations I've built:

A full agent pipeline using Claude Code + n8n that monitors Gmail via Pub/Sub, scores inbound leads with Anthropic's API, and auto-generates personalized proposals. It runs live and handles edge cases.
A GHL CRM MCP server (228 tools) that connects GoHighLevel's entire API surface to Claude-based agents — real-time contact management, opportunity tracking, and conversation automation.
An autonomous Upwork job pipeline: Gmail Pub/Sub push trigger → LLM job scoring → dynamic proposal generation → browser automation submission. End-to-end, no human in the loop.

Most advanced system I've architected: CRESyncFlow — a real estate data sync platform with MCP Apps for interactive UI rendering inside Claude, Anthropic OAuth integration, and live CRM data flows. Production-deployed, multi-tenant.

Claude Code experience: I work with Claude Code daily. I've built multi-agent swarms using Claude's API (spawning sub-agents with scoped tool access), integrated MCP servers for real-time context injection, and deployed in production pipelines where the agent commits code, runs tests, and self-corrects on failure.

60-Day Roadmap (full answer in Q4 below).

If you're building serious internal automation infrastructure for 100+ people, I can get you there. Portfolio: https://portfolio.mcpengage.com

Q1: Most complex automation

[Upwork pipeline — Gmail Pub/Sub → scoring → proposal gen → Playwright submission. 3 prod bugs described.]

Q2: AI task routing architecture

[Intake → Classification (Anthropic API) → Routing (n8n + PM API) → Reporting (Supabase) → Slack UX with action buttons]

Q3: Make.com/n8n vs Claude Code agents

[Deterministic orchestrators vs goal+tools agents. Real power = combining both.]

Q4: 60-day roadmap

[Weeks 1-2: Mental model + first win. Weeks 3-6: Claude Code + production deployments. Weeks 7-8: Architecture + Phase 2 kickoff.]

Q5: Recent similar experience

[Upwork pipeline, GHL MCP server 228 tools, CRESyncFlow, AI coaching for 3 founders, 70+ MCP demo portfolio]

3.2 KiB Raw Permalink Blame History