01
Start and build
Pressure-test an idea, research a market, organize a launch, and make entrepreneurship less lonely.
AgentX AI CourseLive in person · Tenafly, New Jersey
Build your second-brain. Work with superhuman reach.
Stop asking AI one-off questions. Learn to direct a personal agent that can research, organize, plan, remember, and help move real work forward—with you in control.
- Dates
- October 5–27, 2026
- Schedule
- Nine live sessions · 6:30–9:00 PM
- Experience
- 22.5 live hours · No coding required
9 live sessions
22.5 guided hours
1 working personal assistant
0 coding required
Different lives. Same leverage.
AI becomes powerful when it meets the life you already have.
The course starts with your real goals. You learn a method that can travel from a Monday-morning task to a once-in-a-decade decision.
02
A personal assistant that learns your operating system.
Turn recurring context, preferences, projects, and approved memory into a second-brain you can reach when you need it.
- Plan the week
- Draft and revise
- Track open loops
- Find what matters
03
Make the big call
Compare large purchases, model a budget, organize portfolio research, and prepare better questions for a financial professional.
04
Put life in order
Organize documents and photos, plan travel, find discounts and club access, and keep family details easier to retrieve.
01/ 04Swipe or use the arrows
Money clarity
Build a budget, compare scenarios, organize finances, and structure portfolio research without outsourcing judgment or trades.
Important research
Research doctors and medical decisions, gather sources, and organize questions for qualified professionals—not diagnosis.
Better choices
Compare cars and other large purchases, plan travel, shortlist places to go out, and find useful discounts.
Life archive
Create a durable way to organize documents, photos, preferences, and the details your future self will want.
Agentic AI fundamentals
Learn the language. Then build the system.
You learn what makes an agent different from a chatbot, how the parts work together, and where human judgment belongs. Then you make each part useful.
Explore the buildSeven ideas.
One working system.
The nine sessions move from the language engine to a personal assistant you can direct, check, and control. No coding knowledge is assumed.
- Large language model (LLM)Sessions 1–3
- The engine that understands and generates language from your instructions and context.
- AgentSession 4
- An LLM working toward a goal in a loop: plan, act, check, and adjust.
- Agentic systemSessions 4–5
- The full setup around an agent: boundaries, tools, memory, evidence, and approval.
- ToolsSessions 5 & 8
- Approved ways for an agent to search, read files, use software, or create an output.
- MemorySessions 6–7
- Context you choose to keep so the system can recall preferences and prior work.
- EvidenceSessions 4 & 9
- Sources and checks that help you verify a claim before relying on it.
- Human approvalEvery session
- The final gate: you decide what is saved, shared, or used.
- 01Name the goalStart with a real decision, project, or recurring task.
- 02Give useful contextAdd only what changes the answer.
- 03Connect approved toolsDecide what the agent may reach and do.
- 04Check the evidenceFind weak claims before they become decisions.
- 05Remember with permissionKeep what helps. Correct what changed.
- 06You approve the resultNothing important leaves without human judgment.
Four weeks · Nine sessions
Learn it on Monday. Use it by Tuesday.
Every evening has a visible finish line. The course compounds from a first safe experiment to a personal assistant with approved memory and a reachable interface.
Week 01
Make AI useful
Oct 5–6Session 1Meet your future assistant
Create a safe first experiment and see the four-week path from chat to a directed agent.
Session 2Turn a request into a result
Build and revise a finished artifact, then identify what made the work better.
Week 02
Prompt, check, activate
Oct 12–14Session 3Useful prompting and multimodal work
Use goal, context, constraints, and desired output; then turn visual inputs into checked action.
Session 4Find it, check it, fix it
Verify claims and map the parts of an agent: decisions, tools, memory, permissions, and stopping conditions.
Session 5Give intelligence a place to live
Build your Personal Agent Map, activate Hermes Desktop, and get a personalized local response.
Week 03
Remember safely
Oct 19–20Session 6Memory foundations
Connect Hindsight, store one safe memory, test recall, and see what durable context changes.
Session 7Remember, recall, correct
Test memory, correct one miss, and connect approved knowledge to your Agent Map.
Week 04
Reach and extend
Oct 26–27Session 8Reach your assistant from Slack
Connect local Hermes to Slack, send the first message, and practice the recovery lane.
Session 9Prove it, then plan what is next
Complete a conversation and memory capstone, test retrieval, and leave with a departure plan.
Every session6:30–9:00 PM
Course datesOctober 5–27, 2026
FormatIn person · Tenafly, New Jersey
Leave with something real
The work does not disappear when class ends.
A working personal assistant
Configured around context you approve, with a clear path for extending it.
Your Personal Agent Map
A one-page model of the goal, inputs, tools, evidence, outputs, memory, and permission boundaries.
A verification method
A repeatable way to separate recommendations from claims, check sources, and fix weak work.
A first useful result
One practical mission moved from rough intent to a reviewed outcome you can use.
Safety and recovery paths
A permissions checklist and clear way to correct memory, tools, or connections when something misses.
A next-30-days plan
A focused route for turning the course build into an everyday practice.
Executive judgment. Builder credibility.
Learn with someone who understands both what AI can do and what it must not do.
Adam Behrman is an executive advisor on AI, automation, operating-model transformation, value realization, risk, and controls—and a hands-on builder of working AI assistants and agent workflows.
- Former McKinsey banking and advanced analytics leader
- Former Chief Model Risk Officer at a $30B bank
- Advisor to major global financial institutions and senior executives
- NYU Stern MBA · CFA · FRM
- Executive AI study at MIT · AWS Certified Machine Learning — Specialty
AgentX AI Course is an independent program. No university affiliation, accreditation, or transferable academic credit.
October cohort enrollment
Four weeks can change how you work with information for years.
One program. One high-touch cohort limited to 20 participants. The curriculum stays the same; only the enrollment date changes tuition.
Launch
$1,950Enroll through Sep 13$86.67 per live hourStandard
$2,250Sep 14–27$100 per live hourFull
$2,500From Sep 28$111.11 per live hourLaunch tuition is available now.
Tuition includes nine live sessions, the preflight clinic, course materials, guided setup, recovery guidance, and the complete take-home system.
Built for ambitious beginners
You are ready if you can use the laptop you already own.
No prior AI, programming, or command-line experience is required. You should be comfortable with browser tabs, uploads, downloads, copy and paste, and basic files.
Bring
- A Mac with an M1, M2, M3, M4, or newer M-series chip, or a Windows 10/11 laptop
- Permission to install software
- Comfort with browser tabs and basic files
- One real goal you want AI to help move
Mac check: open the Apple menu and choose About This Mac. A supported Mac shows a Chip named M1, M2, M3, M4, or a newer M-series model. Macs showing an Intel processor are not supported. Windows 10 and 11 laptops are supported. A preflight clinic helps finish setup before Session 1.
The practical questions
Before you join.
Is this for complete beginners?
Yes. The course begins with an ordinary AI task and builds one capability at a time. You need everyday laptop comfort, not coding experience.
What exactly will I build?
You will create a Personal Agent Map, configure a local assistant, add approved memory, connect it to Slack, complete one practical mission, and leave with verification and permissions methods.
Can it help with financial or medical choices?
It can help organize research, compare scenarios, surface sources, and prepare better questions. It does not replace a qualified financial, legal, or medical professional, diagnose conditions, or execute portfolio decisions.
Where is the course held?
In person in Tenafly, New Jersey. The exact venue and arrival details are shared with enrolled learners.
What software do I need?
Bring the supported laptop described above with permission to install software. Setup for Hermes Desktop, Hindsight, and Slack is guided during preflight and class.
How does enrollment work?
Review the cohort details, complete secure payment through Stripe, and receive the onboarding and preflight steps by email.
Your next operating system