The five-step sequence before automating a process
When discussing AI implementation, teams often ask what they can automate. Elon Musk’s five-step algorithm suggests starting with the work itself. Question requirements, remove unnecessary process elements, simplify what remains, accelerate the cycle and automate last.
This is a practical discipline for process improvement, rather than a complex theoretical model. It helps avoid investing effort in work that should be removed. It is not a universal guarantee of results. The application to business processes below is LeanAI’s interpretation.
Question the requirements
Processes accumulate rules, checks, reports and approvals whose origins are no longer clear. They continue because someone once decided they were necessary, or because that is how things have always been done.
Every requirement should have a clear purpose and an accountable person. Requirements created by experienced people can still be mistaken. Ask what problem each requirement solves and what would happen if it were removed.
Questioning a requirement means understanding its purpose and justification. Controls needed for safety, quality or data protection must be assessed before the process changes.
Remove unnecessary work before optimising it
Organisations often add another check, spreadsheet or approval. The second step asks whether a process element can be removed altogether, before trying to perform it more efficiently.
Musk emphasises that occasionally having to restore removed elements can show that a team is truly challenging their necessity. In a business process, treat this as a controlled experiment: agree what to monitor and when to restore the step. Remove work that creates no value while preserving process reliability.
Simplify what remains
Optimisation follows removal. Otherwise a team may efficiently improve a process element that should disappear a few weeks later.
Simplification can mean fewer steps, fewer handoffs, fewer data fields, clearer ownership, a single information source or a shorter decision path. A simpler process is easier to understand, manage and eventually automate.
Shorten the whole process, not just an individual task
Accelerating a cycle makes sense after the process has been cleaned up and simplified. Otherwise the organisation simply performs unnecessary work faster.
Look beyond an individual employee’s speed. Much of the elapsed time can sit between tasks, waiting for information, a decision, approval or the next stage. Shortening the entire cycle can matter more than making one task faster.
Automate last
Automation comes fifth. This is particularly relevant to business process automation today. AI can help prepare documents, analyse data, support customers or retrieve information. Its suitability depends on the task, data quality and how outputs are checked.
Technology can also create a trap. Automating a poorly designed process may preserve its underlying problem. Before building an AI agent or integration, ask whether the work needs to be performed in that form at all.
What does this mean for AI implementation in companies?
Technology projects can reverse the sequence: find an interesting tool first, then search for somewhere to apply it. The five-step approach starts with the work and chooses the technology later.
Before starting an AI project, ask whether requirements are justified, what can be removed, what can be simplified and how to shorten the full cycle. Then decide what to automate.
Illustrative example: an employee prepares a weekly report and the team considers automating its production. First ask who uses it, which decision it informs and whether all the data is needed. A short exception summary may be enough, or the report may no longer be needed. This is a hypothetical example, not a LeanAI client result.
This approach helps avoid choosing a technical solution before defining a relevant problem. AI makes it easier to automate an unsuitable process, which makes the order of decisions particularly important.
How LeanAI connects process improvement and AI
The sequence aligns with several LeanAI principles: start with the problem and process, remove work that creates no value, and use AI where it can strengthen a better-designed way of working.
We begin with the leader’s key business challenge and the process constraining performance. The IMPACT method helps identify the priority, understand causes, test a solution, embed a new way of working and review its impact. Automation is one possible tool, rather than a predetermined answer.
LeanAI adds business prioritisation, measurement and sustained organisational change to this practical logic. Before making work faster, establish that it is work worth doing. Then consider how AI could help perform it better.
Common questions about business process automation and AI
Where should business process automation begin? With a business goal and process analysis: identify the constraint, check which steps are necessary, simplify the workflow and agree success measures. Select automation tools afterwards.
Do Lean process improvement and AI replace each other? They can complement each other. Lean helps clarify value, workflow and unnecessary steps. AI can help process information and perform some tasks within a redesigned process.
When is AI automation appropriate? When the task is necessary, the process is sufficiently clear, suitable data is available and outputs can be checked. Test a limited scope and assess errors, costs and the effect on the whole process.
Which challenge most limits your business results?
The first step is a meeting with the leader to understand the current situation and expectations.
Let’s discuss your business challenges →AI and Lean implementation and practical corporate training →IMPACT method →