Use AI tools. Manage them — like five new hires.
A management-grade program for business owners and their teams. By the end, your people don't operate AI — they direct it. Each one is the Manager of Five.
80% of employees stop using AI within three weeks.
/ the-problemMicrosoft tracked 300,000 employees on Copilot. Excitement peaked the first three weeks, then a crater of disappointment, after which most quietly stopped. The technology isn't the problem. The training is.
The 401 trap
Most "AI training" is either 101 — basics ("what is ChatGPT, how do I write a prompt") or 401 — technical ("how do I fine-tune a model, how do I build an agent").
101 produces party tricks. 401 intimidates. Neither produces ROI.
The missing 201
Productivity lives in the missing middle — the 201. The level where people learn to direct AI, not just operate it.
This is the level no one teaches. This is the entire premise of You are the Manager of Five.
The three mistakes that kill AI inside a company.
/ anti-patternsThese are not opinions. They are the consistent failure modes documented across enterprise AI rollouts. Almost every company that fails with AI is making at least one of them. Many are making all three.
Treating AI like a tool, not a collaborator.
People type a question into a chat box, get a mediocre answer, and conclude "AI isn't there yet." They didn't give context. They didn't review the output. They didn't iterate. They treated it like Google.
AI is not Google. It's a collaborator with one day of experience at your company. You wouldn't hire someone and expect quality work without a briefing — yet that's how 80% of users approach AI.
Skipping the 201 — putting people in 101 or 401.
Most corporate AI training falls into one of two failure categories. 101 is too basic to produce real output. 401 is so technical it intimidates and excludes the people who'd actually benefit most.
Productivity lives in the 201 — the missing middle. That's where this program operates. If your training skipped the middle, your team learned the wrong thing.
Treating AI as an IT/security problem, not a management problem.
When IT owns AI, predictable things happen: tools get blocked for "compliance," no one gives explicit permission, and no business owner takes responsibility for outcomes. Employees suspect a trap and keep working the old way.
AI is a management problem. Who decides what to automate? Who trains the team? Who measures impact? If your CIO is handling this alone, you've already lost.
The six skills that distinguish the 20% who succeed.
/ the-sixAcross the research, the same six skills show up in everyone who actually gets value out of AI. The course is structured so each one gets developed in depth.
Context Assembly
Knowing what background, constraints, and information to provide before asking. The single biggest determinant of output quality.
Quality Judgment
Knowing when to trust the output and when to verify. Calibrated trust, not blind acceptance, not blanket skepticism.
Task Decomposition
Breaking work into AI-appropriate chunks vs. human-appropriate chunks. The art of knowing what to delegate.
Iterative Refinement
Treating the first draft as a starting point, not a finished product. Most failures come from accepting v1 as final.
Workflow Integration
Embedding AI into existing processes rather than treating it as a side tool. Where the productivity actually compounds.
Frontier Recognition
Knowing explicitly when to stop using AI for a task. The "jagged frontier" punishes those who don't know its edges.
The three levers — only the CEO can pull these.
/ implementation-modelTraining your team is necessary but not sufficient. There are three organizational decisions that no consultant, no IT department, and no manager-of-managers can make for you. They have to come from the top.
Create an AI Lab — with non-technical staff
The biggest beneficiaries of AI are not engineers. They're operations, sales, finance, customer success. But they will not discover AI on their own. Create a formal space — 1–2 hours weekly, or a monthly sprint — where non-technical staff bring real workflows and rebuild them with AI. The experiment: "this 4-hour task I do today — what does it look like with AI?"
Systematize sharing failures
This is counterintuitive, which is why almost no one does it. Most companies share success stories. Companies that win with AI share failures: "I tried to make AI do X and it got it wrong this way." Without this, the jagged frontier silently makes your team's work worse, and no one notices. A living failures page is more valuable than a hundred best-practice documents.
Pull AI out of IT — and grant explicit permission
Most companies have AI deployed technically, but no one ever told the employee: "Yes, you're allowed to use this. No one will be punished. We want you to use it." Without explicit permission from leadership, employees suspect a trap and keep working the old way. The permission must come from the CEO or direct manager — not from an IT email.
The framework: You are the Manager of Five.
/ the-premiseBefore any module, the audience needs to absorb the reframe that makes everything else click.
The old role
An expert operator. You knew the business, you knew the tools, you executed the work. Your value was the doing. Most of your day was spent producing output yourself.
The new role
You are the Manager of Five. Five Harvard-grade collaborators do the doing. Your job is briefing, mission, outcome, review. Your only irreplaceable contribution is context. The course teaches you how to externalize it.
Two products, one program.
/ architectureThe program is delivered as two complementary products. The Executive Briefing sells the change to the person who decides. The course teaches the people who execute. They are designed to be deployed together.
Executive Briefing
Audience: the CEO or business owner — the person who signs the check, the person who decides whether to bring AI into the company.
Outcome: they understand why typical AI training fails, what role they personally must play (not delegate to IT), and they approve the rollout for their team.
You are the Manager of Five (Course)
Audience: the operators, managers, and individual contributors who will use AI day-to-day.
Outcome: operational fluency. They leave knowing how to direct five AI collaborators, build harnesses around them, and integrate them into existing workflows.
Three modules. Each one builds on the last.
/ modulesThe course is structured so each module answers a question the previous one set up. By the end of Module 3, the audience has seen the framework, the mental model, and five real cases pulled from a working business.
The Briefing
Why most AI efforts fail. The three mistakes. The six skills that separate the 20% from the 80%. The implementation model. The course promise.
The Harness
The horse-and-harness analogy. The seven pieces that convert raw chat into directed work: memory, context, tools, skills, connections, triggers, channels.
The Harness in Action
Five real cases from DiabetesDME, in escalating complexity. Each one mapped to harness pieces. The MCP case as the climax — the cherry on top.
The Briefing — why you failed before, and why this time is different.
/ module-01Objective. Leave the audience with clarity on three things: (1) why most AI efforts fail, (2) the six skills that separate the 20% from the 80%, and (3) the organizational model that makes the difference.
The statistic (5 min)
The 80% drop-off after three weeks. Microsoft's 300,000-employee Copilot study. Why this happens to almost every rollout — including the well-intentioned ones.
The three mistakes (15 min)
Same three mistakes from the executive briefing — but framed for the operator: "If you treat it like Google, you'll abandon it. If your training was 101 or 401, you got the wrong training. If you're waiting for IT to tell you what to do, you'll wait forever."
The six skills (20 min)
Each skill introduced with a one-line definition and a concrete example. They will be developed in depth across Modules 2 and 3 — but the audience leaves Module 1 knowing the names.
The implementation model (15 min)
The three levers, framed for the team: "Even if your CEO doesn't pull these levers, here's how you can pull mini-versions of them yourself in your own corner of the company."
The promise (5 min)
"By the end of this course, you are the Manager of Five." Set the bar. Set the stakes. Open the path.
The Harness — what to attach to the chat so it does work.
/ module-02Objective. Make the audience understand that ChatGPT or Claude alone is like a horse without a harness. Strong, fast, intelligent — but tied to a cart with no harness, it pulls nothing. The harness is what converts model power into directed work.
The seven pieces of the harness.
Memory
"How do I avoid repeating everything every time?"
Files, folders, Obsidian, glossaries, documented business rules. Persistent context that doesn't evaporate at the end of the chat.
Context
"How do I explain my business so the answers are useful?"
Role briefings, customer data, history, constraints, what good looks like. The single biggest lever on output quality.
Tools
"How do I give it specific capabilities — calculate, search, read a PDF?"
Tool use: web search, calculator, file reader, code interpreter. The verbs the model can perform on your behalf.
Skills
"How do I teach it a process once so it executes when the input arrives?"
Documented recipes that fire on a trigger. Skills are how a one-time conversation becomes a reusable capability.
Connections (MCPs)
"How do I give it live access to my systems — Zoho, Slack, calendar?"
MCP servers, plugins, connectors. The shift from uploading files to granting access. This is the qualitative leap.
Triggers / Schedules
"How do I make it work without me being there?"
Scheduled tasks, event triggers, automations. The transition from synchronous chat to asynchronous agent.
Channels & Artifacts
"How do I communicate with it — and how does it return work?"
Slack, email, live artifacts, persistent dashboards. The interface layer between the operator and the agent.
The harness reflex
"Every time AI gives you a mediocre result, it's not that AI is bad — it's that a piece of the harness is missing."
Replace the old reflex of "AI doesn't work for this" with the new reflex: "which harness piece am I missing?"
The Harness in Action — five real cases from a working business.
/ module-03Objective. Show the harness applied in escalating complexity. Each case follows the same template — Pain Point → Harness pieces attached → Result. The audience leaves with a reproducible mental pattern they can apply to their own business.
The case template.
pay-gabriela — a documented recipe.pay-call-center.cfo-balance / ddme-cfo.ddme-claim-items.ddme-claim-items, already built.The harness pieces — built case by case.
Here is the entire module on a single screen. The harness is built piece by piece. You don't need to reach Case 5 tomorrow — you need to start at Case 1 today.
| case | tool | context | skill | mcp |
|---|---|---|---|---|
| 01 · Gabriela | ✓ | ✓ | ✓ | — |
| 02 · Pakistan | ✓ | ✓ | ✓ | — |
| 03 · True Balance | ✓ | ✓✓ multi-source | ✓ | — |
| 04 · AR Dashboard | ✓ | ✓✓✓ business rules | ✓ | — |
| 05 · Sales · MCP | ✓ | ✓✓✓ | ✓ | ✓ |
What comes after Module 3.
/ roadmapModules 1–3 are the core. They are enough to produce real productivity gains across an organization. Future modules go further — for teams that want to push past the basics.
Skill Building
How to document your own processes so AI executes them. Going from "I have to explain it every time" to "the skill fires when the input arrives."
Agent Operations
Triggers, schedules, multi-channel agents. The shift from synchronous chat to AI that runs while you sleep.
Decision Framework
What to automate, what to keep human, where AI degrades performance (the jagged frontier). The matrix every Manager of Five carries in their head.
Risk & Governance
Quality control at scale. Verification protocols. Where to insert a human review. How to supervise five collaborators who work faster than you.
Implementation Plan
The 30/60/90-day rollout for your team. Specific milestones, specific outputs, specific anti-patterns to watch for.
Build Your First Skill
Hands-on: every participant ships one working skill that automates a real process from their week. The proof that the program landed.
One conversation is enough to know.
Twenty minutes. We talk about your team, your goals, and whether You are the Manager of Five is the right fit. No pitch.
Book a free 20-min consult →