ai curriculum, 2026
Agentic workflows
10 modules. 65 lessons. 4+ hours.
The system
Six protocols.One system.
Every protocol teaches you how to build one part of your AI system. None of them stand alone. Each one feeds the next, and together they form one connected machine that runs your business.
- Protocols
- 6
- Modules
- 18
- Days
- 30
01 IntelligenceTeach your AI how to think, write, and prioritize for your business.
Generate personalized outreach scripts, build a daily lead priority list, and produce market update content using AI, without writing anything from scratch. The output sounds like you, not a generic script.
Why this comes first
If the scripts are generic, the calls don't convert. This is the brain of the entire system. Get it right first and everything downstream improves automatically.
What you'll learn
- Write prompts that get usable output on the first try
- Build an agent profile that gives your AI context
- Create a script library for expireds, FSBOs, leads, neighbors, and SOI
- Set up a daily lead priority list from your CRM activity
- Generate follow-up sequences and appointment briefs
02 OutreachYour AI makes the calls so you don't have to.
A live AI voice outreach campaign running on its own. Expired listings and FSBOs called automatically every morning. New leads receive a call within 90 seconds, day or night. Every outcome logged in your CRM.
Why this comes here
The reason most agents don't call expireds every day is consistency, not motivation. This protocol removes you from the prospecting equation entirely.
What you'll learn
- Launch your expired listing campaign
- Launch your FSBO campaign and speed-to-lead trigger
- Set up neighbor prospecting on every listing and closing
- Review call logs and adjust scripts based on conversions
03 Follow-UpNo conversation ever dies from silence.
Every lead that doesn't book immediately enters a structured follow-up sequence. Texts, emails, and voice calls go out automatically. Your sphere hears from you four times a year, without you writing a single message.
Why this comes here
Most agents stop following up after the second or third attempt. That's where the deal dies. This protocol makes sure you never disappear from a lead's radar.
What you'll learn
- Build a 14-day multi-touch follow-up sequence
- Set up SOI nurture on a quarterly schedule
- Build a reactivation campaign for dormant contacts
- Load the price reduction conversation handler at day 21
04 CaptureTurn every response into a booked appointment and clean CRM data.
Every lead who responds goes through an interactive video experience that collects their information and sends it directly into your CRM. No manual data entry. You walk into every listing appointment already knowing what the seller is thinking.
Why this comes here
Getting a lead to respond is only half the job. If you don't have a clean way to capture their information, you lose them. This protocol closes that gap permanently.
What you'll learn
- Build a home valuation capture funnel
- Build a pre-listing intake that collects seller expectations
- Create a buyer pre-qualification flow
- Build a post-close testimonial collector
- Connect every funnel to your CRM automatically
05 OperationsThe infrastructure that holds everything together and keeps it running.
Your CRM fully configured, every tool connected, leads routing to the right place automatically, and a clear view of your entire pipeline at any moment. Nothing sitting in a gap between systems. Nothing requiring a manual handoff.
Why this comes here
Five tools that aren't connected are just five tools. This protocol turns everything built in Protocols 1 through 4 into one machine.
What you'll learn
- Set up your CRM with pipeline stages, tags, and dashboards
- Build the five core automation connections every agent needs
- Set up lead routing, agent notifications, and after-hours handling
- Run a weekly review habit that keeps the system healthy
06 GrowthEvery business event becomes an automatic growth action.
Every listing you take, every price reduction, and every deal you close automatically triggers a growth campaign. You won't be doing more to grow. The system uses what's already happening to create new opportunities on its own.
Why this comes here
Most agents treat a new listing as just a listing. This protocol treats it as a growth trigger. Every listing means buyers who need to know. Every closing means a referral opportunity.
What you'll learn
- Build a new listing buyer campaign
- Set up price reduction buyer alerts
- Build an AI content package generator for every listing
- Create a monthly market update campaign for long-cycle leads
- Run a monthly performance analysis using CRM data and AI
How it works
Ourcapabilities.
Master Claude Code
Learn the full interface, permission modes, and daily slash commands that professionals use to ship AI-powered products faster.
Read moreBuild Real Products
Go from spec to deployed website with competitor intel, structured design briefs, and parallel agent builds that actually work.
Read moreScale with AI Systems
Wire up memory, MCP servers, hooks, and subagents to automate your entire workflow and build a system that compounds.
Read more
The curriculum
A structured systemfor shippingAI-powered products.
Built for developers who want to ship real products, not just understand theory. Every module is practical, every lesson is actionable.
Parts
10 Core Modules
Foundation to revenue, every module builds on the last.
Lessons
65 Focused Lessons
No filler. Every lesson ships a real skill.
Hours
4+ Hours
Dense, structured, high-signal content.
Bonus
11 Freebies
Production files, templates, and configs.
Latest modules
Thesyllabus.
01 Foundation + SetupInstall → CLAUDE.md → Plan Mode
1.1
Install Claude Code
- Desktop app download
- Terminal install path
- First-run + API key
1.2
The interface tour
- Sidebar + settings
- Terminal vs GUI
- Antigravity vs VS Code
1.3
Free vs Pro vs Max
- What each plan gets you
- Which one you actually need
- When to upgrade
1.4
Permission modes
- Before Edits (safe default)
- Edit Automatically (speed)
- Bypass (live builds only)
1.5
The 5 Claude surfaces
- Chat · Cowork · Claude Code
- Managed Agents · API
- When to reach for which
1.6
Daily slash commands
- /context /compact /clear
- /model /cost
- Decision tree for each
1.7
Organize your desktop
- Ask Claude what you work on
- Auto-create folder structure
- 30 seconds, done
1.8
Connect your tools
- Supabase, Gmail, Pinecone
- YouTube, Notion, Drive
- When to build a custom connector
1.9
Folder structure
- Hard-coded folders: Code vs Cowork
- Where Skills live
- Where CLAUDE.md goes
1.10
Git + GitHub backup
- Why you need it from day 1
- 2-minute setup
- Bitwarden for secrets
1.11
CLAUDE.md + Plan Mode
- The two biggest unlocks
- Walk through my production CLAUDE.md
- Plan Mode live demo
02 Build a Real WebsiteCompetitor Intel → Design → Host
2.1
The 3-step website system
- Competitor Intel → Design → Host
- Why spec-first always wins
- What we're building today
2.2
Scraping competitors
- Firecrawl + Apify
- Capture structure + copy + styling
- Turn scrape into structured output
2.3
Structured design brief
- Research → spec
- What Claude needs
- The brief template
2.4
UI UX Pro Max Skill
- What this Skill does
- Live run
- Yours to keep
2.5
Parallel agents
- Build 3 versions at once
- Compare side-by-side
- Pick and iterate
2.6
SEO baked in
- Meta + schema markup
- Sitemap + robots.txt
- Core Web Vitals
2.7
UI Sniping
- Steal proven design patterns
- Which sites to study
- How to adapt
2.8
Deploy to Vercel
- Connect the repo
- Environment variables
- Live URL on camera
2.9
Verify Claude's work
- The #1 beginner mistake
- What to check
- How to reject bad output
2.10
Debug like a pro
- Read the error first
- Git stash + reset
- Re-prompt with context
03 Power FeaturesSkills · Hooks · MCP · Subagents · Critic Loop
3.1
Skills from scratch
- What Skills actually are
- The SKILL.md structure
- Your first working Skill
3.2
Skill marketplaces + security
- Cowork + skills.sh
- What to install
- Security vetting
3.3
Build your own Skill
- Live walkthrough
- Reusable patterns
- Yours to keep
3.4
Token management
- What things cost
- Model routing
- Claude vs Codex efficiency
3.5
Context rot explained
- When conversations go bad
- /compact vs /clear
- Short focused sessions
3.6
Routines (24/7 agents)
- What Routines are
- Building your first one
- Real use cases
3.7
Hooks automation
- Pre/post tool calls
- Auto-format, auto-test
- Contact form demo
3.8
MCP in plain English
- What MCP is
- Pre-built vs custom
- Why you'd need one
3.9
Build a custom MCP
- Live server build
- Deploy + share
- Wire to Claude Code
3.10
Subagents + worktrees
- Delegate focused work
- Parallel builds
- When to use each
3.11
Autonomous critic loop
- LLM-as-judge pattern
- Self-correcting output
- Advanced orchestration
3.12
Dispatch · Scheduled · Plugins
- Dispatch (newer way)
- Scheduled Tasks
- Plugin marketplace
04 Memory SystemPinecone · NotebookLM · Granola · Obsidian RAG
4.1
The 3 levels of memory
- Context (right now)
- Project (per-project)
- Long-term (forever)
4.2
NotebookLM
- Knowledge that persists
- Query everything you've made
- Audio overviews
4.3
Pinecone
- Vector memory, infinite recall
- Wire to Claude Code
- Real queries on camera
4.4
Granola meetings
- Auto-transcribe every call
- Turn meetings into memory
- Query what was said
4.5
CSV folder memory
- Structured long-term data
- Claude queries directly
- When this beats Pinecone
4.6
Obsidian RAG (Karpathy)
- Vault-as-second-brain, local
- Embed notes, query with Claude
- Karpathy workflow end-to-end
05 Hermes AgentBuild your own AI operator and wire it to memory
5.1
Build your Hermes agent
- What a Hermes agent is
- Why you'd want one
- How to build it
5.2
Connect to memory system
- Wire Hermes to Pinecone
- Unified interface
- Query anything across sources
06 AppsBLAST framework + Stripe + auth + Resend
6.1
App architecture
- Front-end
- Back-end (DB, auth, APIs)
- Automations + integrations
6.2
BLAST Framework
- What BLAST stands for
- When to use it
- Live walkthrough
6.3
Sign-in flows
- Modal + Claude Routines
- Auth that works
- User management
6.4
Resend email marketing
- Wire up Resend
- Transactional + broadcast
- Capture leads · trigger sequences
6.5
Stripe payments
- Wire up Stripe
- Live payment on camera
- Handle edge cases
07 Build AnythingFirst principles · lead-gen pipeline · agency stack
7.1
First-principles system design
- The Musk framework
- Break any problem down
- Rebuild systematically
7.2
Lead scraping → Instantly
- Apify + Firecrawl
- Enrich data via Claude
- Push to Instantly
7.3
Marketing agency stack
- Full agency setup
- Client delivery flow
- Packaging for scale
08 Design SystemsHTML decks · 3D · video · images · logos
8.1
Claude Code Design
- The brand-new design feature
- vs Figma / Stitch
- Live design walkthrough
8.2
HTML presentations
- Animations baked in
- Export and share
- Like this workbook
8.3
Animated graphics
- Drop-in for any website
- CSS-only + GSAP
- Motion that converts
8.4
Spline 3D
- When 3D is worth it
- Performance tradeoffs
- Integration walkthrough
8.5
ElevenLabs + Higgsfield
- Voice generation
- Video generation
- Combined pipeline
8.6
Nano Banana images
- Best prompt patterns
- Production workflow
- Quality control
8.7
Auto-fetch brand assets
- Scrape any logo from any URL
- Live asset library
- Drop into builds
09 Compliance & MaintenanceHIPAA · GDPR · API hygiene · monitoring · SDK
9.1
HIPAA · GDPR · API keys
- What not to share with Claude
- Self-hosting options
- API key hygiene
9.2
Safer build prompts
- Prompts that catch bugs
- Security audit workflow
- Before you ship
9.3
Must-have stack
- RAG · Claude · Antigravity
- Vercel · GitHub
- What each piece does
9.4
Monitor in production
- What to actually track
- Error recovery
- Uptime + alerts
9.5
SDK + certification
- Anthropic SDK basics
- The certification path
- How to stay current
10 Turn This Into $$$Pricing · packaging · proposals · Build & Sell
10.1
Principles that save 5 years
- Mindset shifts
- What actually matters
- Patterns I'd steal first
10.2
$30K to $50K deals, zero outbound
- How I land them
- Positioning that works
- The follow-up system
10.3
The Build & Sell system
- My actual playbook
- Client delivery + invoicing
- Pricing framework
10.4
Where to go next
- Skool community
- Weekly live builds
- How to keep learning
11 freebies
Bonusfiles.
links go live when videos publish
01 / Template
CLAUDE.md Production Template
The exact file used in every build. Copy, fill in context, run.
02 / Skills
5 Production Skills
Ready-to-install for the five most-used workflows. Drop in and go.
03 / Automation
Hook Configurations
Pre-built pre/post tool call hooks. Auto-format, auto-test.
04 / Server
MCP Server Starter
A working MCP server to fork and deploy. Skip the boilerplate.
05 / Memory
Pinecone Memory Starter
Pre-wired memory config. Connect once, query forever.
06 / App
BLAST App Template
The base repo for every shipped app. Auth, DB, API pre-wired.
07 / Pipeline
Lead-Gen Pipeline Config
Apify + Firecrawl + Instantly. One pipeline out of the box.
08 / Design
Design Brief Template
The structured spec handed to Claude before every build.
09 / Ops
Compliance Checklist
HIPAA + GDPR + API hygiene in one doc. Print it, pin it.
10 / Prompts
Prompt Library
30+ production prompts by use case. Copy, paste, ship.
11 / Community
Private Community Access
Weekly live builds, direct Q&A, early access to new modules.
Your instructor
Meet yourinstructor.
James Rembert. Aigentic AI architect.
James teaches the methodology, then helps your whole team run on it. Agentic workflows are how businesses actually put AI to work. Today James runs a boutique advertising agency in Philadelphia that works exclusively with the real estate industry, while teaching agentic AI methodology to business leaders nationwide.
His edge: RAG agent development, agentic workflow design, and a decade of real-world advertising execution. Not theory. Systems that work.
- Years building
- 10+
- Products shipped
- 50+
- Students taught
- 1K+
Enrollment
Ready to buildwith AI?
see you inside.
10 parts, 65 lessons, 4+ hours of focused, no-filler instruction. Plus 11 production-ready bonus files you can use today.
Join the labskool.com/aigent-lab
10 parts. 65 lessons. 11 bonus.
