How to build business systems so your company runs without you
Six phases and a 12-point checklist, in the order most advice gets backwards: the people and their incentives first, the SOPs fifth, AI last.

On this page · 10 sections
- 01 What does a business system need to run without the owner?
- 02 Why do documentation and AI come last?
- 03 How do you build business systems? The 6 phases
- 04 How do you keep a business system running every week?
- 05 When is a business system done? The 12-point checklist
- 06 What does the owner do once the systems run?
- 07 Which system should you build first?
- 08 Questions people ask
- 09 The bottom line
- Sources
The short answer
To build business systems so the company runs without the owner, give each system one owner and one KPI, pay the people in it toward that KPI, review it every week, and launch it “good enough” before you document it or add AI.
- A business system is people, process and tools that run the same way every time and take the company to one KPI.
- You build it in six phases: define it, communicate and train, launch it, reach 70 to 80% adoption, document it, then improve its efficiency.
- Documentation comes fifth. AI comes last, as a tool inside a process that already works.
- A system is done when it passes 12 checks, including running without the owner’s daily intervention and delivering results for 2 consecutive quarters.
Your team writes documents, runs meetings and tries every new AI tool. Revenue doesn’t move, and the decisions that matter still wait for you. That is rarely a people problem or a tools problem. It’s a missing system. Here is the order I build systems in, the weekly cadence that keeps them alive, and the checklist I use to call one done.
What does a business system need to run without the owner?
A business system needs three pillars to run without the owner: people, process and tools, all cascading down from your company strategy. And it’s all about the incentives: how people move in your team, why they move, and what makes them move.
| Pillar | What it holds | The question it answers |
|---|---|---|
| People | The right people in the right roles, job descriptions they know, economic incentives, and one system owner with a clear reporting line and a single KPI | What do I pay you for? |
| Process | How the information flows, when, and who handles it; who decides (decision rights); where work changes hands (the hand-offs); built for the 90% of cases that get results | Who does what, and who decides? |
| Tools | The cadence (weekly, monthly, quarterly, yearly), a performance dashboard, a taskboard, and the technology that supports the process | How do we see, every week, whether it works? |
It’s close to the people, process, technology framework. Here, tools means the cadence and the dashboards as well as the technology.
What a system is not: a library of SOPs or documents, a wishful idea or strategy, an isolated meeting, a Loom video or a custom GPT, a person, a department, or a tool or software. None of them is a system on its own.
Why do documentation and AI come last?
Documentation and AI come last because neither one builds a system: documents record a system that already runs, and AI speeds up a process that already exists. Much of the advice on building systems starts with writing SOPs. I put documentation fifth.
When I get into a company, what I find is noise. People write hundreds of documents and bury each other in them, and no one knows what a document is for or which KPI it moves. AI raised the activity and the output. It didn’t raise revenue, performance or EBITDA: the numbers that matter. That’s AI workslop, and it’s a system problem.
People are the engine. Software and IT projects fail when people lack a clear why, when training and support are missing in the first 90 days (they decide adoption), when rewards don’t match the behaviors you want, and when there are no feedback loops. That last one is what the cadence is for.
This is not a new idea. Reviewing case studies and company-level data in 2000, the economists Erik Brynjolfsson and Lorin Hitt found that the organizational investments made around computers strongly shape what companies get back from them1.
My rule fits in one line: “The AI tool and the software serve the process that serves the person, and not the other way around.”
How do you build business systems? The 6 phases
You build a business system in six phases: define a “good enough” system, communicate and train, launch it, reach 70 to 80% adoption, document it, then improve its efficiency. Every phase aims at one place, “there”: the system’s main KPI, connected to your P&L.
Define a “good enough” system to get “there.” Its impact on your strategy and market. A single owner. The individual contributors, with job descriptions, incentives and KPI-driven reports. A draft process and good-enough tools.
Communicate and train. The new incentives. The organizational moves: roles, job descriptions, reporting lines. The process and tools, the taskboard, the performance dashboard and the operating cadence (team reviews, 1:1s, quality checks). Then training sessions.
Launch it, good enough. The team cadence, the team performance dashboard and the team taskboard go live. 1:1s help each person adopt it. Feedback from the meetings goes into the owner’s improvement list. Tell the whole company who owns the system: if you want X, talk to the owner.
Reach 70 to 80% adoption: the green light. At that point the system takes you “there,” to the KPI.
Document it. SOPs and videos, added to your company’s library of systems. For a sales system, that document is your sales playbook.
Improve its efficiency. This is where AI and software speed up a process that already works.
How do you keep a business system running every week?
You keep a business system running with its cadence: regular, predictable intervals, weekly, monthly, quarterly and yearly, where the system runs and the team reviews it. The cadence is your operating rhythm, and it’s another driver of accountability. It has three parts:
- The team review: system status, improvements, performance tracking, feedback and the team’s collective intelligence.
- 1:1s: change management. They help each person adopt the new system, iterate with feedback and get coaching.
- The quality check: a self-managed forum, without the owner, where team members check that the system is followed. This is what makes it self-sustaining, predictable and scalable.
The weekly dashboard review is the most important meeting a leader has each week. Ask three questions, across people, process and tools:
- What’s working? What do we do more of?
- What’s not working? What do we do less of?
- What needs to change to get “there,” to the KPI?
The dashboards, by level:
| Level | Dashboard | What it covers |
|---|---|---|
| Leadership | Company performance | Company-wide performance systems, work across departments and organization design. Accountability for revenue and EBITDA. |
| Mid-manager | The team’s system | System improvements, bottlenecks, each person’s performance and the skills the team needs. |
| Individual contributor | The taskboard | Daily focus, priorities and next actions. |
When is a business system done? The 12-point checklist
A business system is done when it passes all 12 points below. I bring AI in only after every box is ticked.
- Single-owned: everyone knows the owner, who to ask for improvements and who holds them accountable.
- One #1 KPI that measures what matters, owned by the owner.
- Economic incentives linked to that KPI: people are paid toward what matters, tied to company strategy.
- Cadence-driven: team and individual cadence in place.
- A team performance dashboard the team uses to decide, prioritize and discuss improvements.
- An individual taskboard each contributor uses to prioritize the day.
- Adopted: more than 70% of the team uses it as designed. No lone wolves.
- Predictable: it runs the same way every time, without constant reinvention.
- Self-sustaining: it runs without the owner’s daily intervention, because the contributors run the quality checks.
- A single communication channel.
- Documented: workflows and roles written down and accessible.
- Delivering results: measurable outcomes tied to business goals for 2 consecutive quarters.
Two points carry the rest. Without pay tied to the KPI (point 3), people move toward something else. Without adoption (point 7), you have a plan, not a system.
What does the owner do once the systems run?
Once the systems run, the owner becomes a system enabler: you improve the systems every week from the cadence and the dashboard, coach and develop the team, and assess people, process and tools. You stop pushing for results and start fixing the piece of the system that blocks them.
| The results-oriented leader fallacy | The system enabler |
|---|---|
| “Why didn’t you hit the number? Push harder. We need more sales. Focus on the results.” | “Why can’t we sell more? What’s blocking us? Is it the process, tools, skills or incentives? Where’s the bottleneck? Let’s fix this piece of the system.” |
| Ineffective and demoralizing | Enabling and empowering |
One rule protects every system: don’t break the ownership line. If someone owns a KPI and has a direct team, no one else in the company asks or reports about that KPI or its process to that team. Delegate ownership, pay people to do their job, give them the resources, do your 1:1s with the owner, and trust.
If you are not sure what still waits for you, the founder bottleneck guide shows how to find it in one week of your calendar.
Which system should you build first?
Build first the system that sits where revenue or EBITDA falls short, and find it by diagnosing before you ask for resources. When the number is short, “sell more” and “we need more people” are the easy answers. Neither one finds the piece of the system that isn’t working.
Diagnose the root cause in people, process and tools:
- Is it distribution or positioning? Leads, discovery calls, pitches or sales enablement? Marketing or sales?
- Where is the bottleneck, and what is the root cause?
- Are the right people in the right place? Do they know what we pay them for?
- Does the weekly cadence reward the right behaviors? Are people clear on their KPIs, and can they measure themselves?
- Is there a single owner? The right tools? Is their time in the right place?
At Digital Smile Design, I built a new business unit as a fractional executive, with a new service offer, a new team, positioning, pitch and sales process, built as systems of people, process and tools. It went from $0 to $200K in sales in one quarter (client cases). I also ran this working session with the leadership team there, and I run it with other leadership teams as a business systems workshop.
Questions people ask
How do I build systems so my company runs without me?
Start with the system your company needs first, “good enough”: one KPI, a single owner, people paid toward that KPI, a draft process and good-enough tools. Communicate and train, then launch it with a weekly cadence. At 70 to 80% adoption, it gets the green light. Document it, then improve its efficiency. It’s done when it passes the 12-point checklist, including running without the owner’s daily intervention and delivering results for 2 consecutive quarters.
Should I write SOPs before I build a system?
No. A business system’s SOPs come in phase 5, after the system runs and 70 to 80% of the team uses it. A library of SOPs is not a system. When no one knows what a document is for or which KPI it moves, it adds noise. Written after the launch, the SOPs and videos record a system that works, and they go into your company’s library of systems.
We rolled out AI and revenue didn’t move. What do we fix first?
Fix the base system first. AI raises activity and output, not revenue, performance or EBITDA. Give the system one owner and one KPI tied to your P&L, pay people toward it and review it every week on one dashboard. Then look inside the process for where AI or software can improve it.
How long does it take to build a business system?
A business system has checkpoints rather than one deadline. The first 90 days decide adoption, so training and support matter most then. Real adoption happens in the first year, through the cadence’s feedback loops. And a system counts as done only once it delivers measurable results tied to business goals for 2 consecutive quarters.
The bottom line
A company runs without its owner when its work runs on systems, not on heroes. Build each one in order: the people and their incentives, a process good enough to launch, a weekly cadence, then the documents, then AI. Call it done only when it passes the checklist.
Build the system, and the results follow.
Sources
- Erik Brynjolfsson and Lorin M. Hitt, “Beyond Computation: Information Technology, Organizational Transformation and Business Performance”, Journal of Economic Perspectives 14, no. 4, Fall 2000, pages 23 to 48. aeaweb.org/articles?id=10.1257/jep.14.4.23




