Almost every company knows the places where work takes longer than it needs to: the quote waiting for a signature, the delivery address typed in twice, the question only one colleague can answer. These places are the best starting point for process optimization, because a small change there often saves a lot of lead time.
This guide explains what process optimization is, how to approach it in five phases, which methods have proven themselves and what it looks like in a small or mid-sized company. It belongs to the topic of AI process documentation, since every optimization starts with a process that has been described.
What Is Process Optimization? A Definition
Process optimization is the planned improvement of an existing workflow, measured against goals set in advance: shorter lead time, fewer errors, less rework, clearer responsibilities or less dependence on individual people. The starting point is always a known current state, and the result is a target state that people actually work by.
Three terms sit close together and are often mixed up:
| Term | What changes | Typical result |
|---|---|---|
| Process optimization | The workflow itself: steps, order, responsibilities, handoffs | Less waiting, fewer loops, clear ownership |
| Process digitization | The medium: the workflow runs in a system instead of on paper | Less re-keying, data in one place |
| Process automation | The execution: steps run without human input | Less manual work on uniform tasks |
In practice the three overlap. A rule of thumb has proven useful: improve the workflow first, then digitize it, then automate it. Automating a detour gives you a faster detour.
For companies with a quality management system, process optimization is also a requirement of the standard. ISO 9001:2015 asks in clause 4.4 for the processes of the QMS to be determined, evaluated and improved, in clause 9.1 for monitoring and measurement, and in clause 10.3 for continual improvement.
The 5 Phases of Process Optimization
Most improvement models, from the PDCA cycle to DMAIC in Six Sigma, follow the same basic logic. For a mid-sized company, five phases are enough.
Phase 1: Capture the as-is process
Write down how the process really runs today: who does what, in which order, in which system, on which piece of paper and with which exceptions? Talk to the people who carry out the workflow every day, and to several of them along the flow, because each one sees only their own part. The guide on as-is process mapping shows how to do this in short conversations. Record your baseline in this phase as well, for example the lead time of a sample of cases, so that phase 5 has something to compare against.
A swimlane diagram helps with the visual side, because it assigns every step to a role and makes handoffs visible. To set the scope beforehand, a SIPOC works well: suppliers, inputs, process, outputs and customers on one page.
Phase 2: Analyze weak spots
With the as-is process in front of you, look for the usual patterns: waiting times, re-keying of data between systems, duplicate work, loops caused by missing information, steps only one person can do, and approvals that check nothing. For each weak spot, note who mentioned it and how often it happens. With a source attached, you can later discuss the size of a problem on facts. The guide on process analysis covers the methods in detail.
Phase 3: Design the target process
For the most important weak spots, design the future workflow. Check the simple levers first: Can a step be dropped? Can two steps run in parallel? Does this approval need a value threshold? Is cover arranged? A RACI matrix clarifies for each step who does it, who is accountable, who is consulted and who is informed. Technical solutions come in where work is left over after the organizational changes.
Phase 4: Implement
Start with one area or one team, collect the special cases the design was still missing, and adjust. Set a date from which the new workflow applies, so the old and the new way do not run side by side for long. Update the process description and the work instruction once the new workflow is in place, because an improvement that is written down nowhere fades with the next change of staff.
Phase 5: Measure and adjust
Compare the new workflow with the starting point, using the metrics you collected in phase 1. A few easy-to-collect figures work well:
| Metric | What it shows | How to collect it |
|---|---|---|
| Lead time | Time from trigger to result | Sample of cases with date of receipt and completion |
| Share of waiting time | How much of the lead time is waiting | Have people estimate the processing time and compare it with the lead time |
| First-time-right rate | Share of cases without a query or correction | Count queries over two weeks |
| Error and complaint rate | Quality of the result | Complaints or correction entries per period |
| Number of handoffs | How often a case changes hands | Count them in the swimlane diagram |
After a few weeks the cycle starts again, with the next process or the next weak spot in the same process.
Process Optimization Methods
The well-known methods do not rule each other out. They set different priorities, and many companies combine them.
Lean management
Lean comes from the Toyota Production System and aligns every process with the value for the customer. Anything that adds nothing to it counts as waste. Taiichi Ohno distinguished seven kinds: overproduction, waiting, transport, overprocessing, inventory, unnecessary motion and defects. In the office they look like reports nobody reads, cases sitting in a tray, forwarding by email, duplicate checks and queries. Lean is a good place to start, because the categories are easy to grasp and apply directly to your own workflow.
Kaizen and continuous improvement
Kaizen is Japanese for change for the better. The idea is known as continuous improvement: many small improvements, suggested by the people who carry out the process, in a fixed rhythm. The usual framework is the PDCA cycle (plan, do, check, act), popularized by W. Edwards Deming. Continuous improvement suits companies that prefer to make one thing better every week over running large projects.
Six Sigma
Six Sigma is data-driven and aims at less variation and fewer defects. It follows the DMAIC cycle: define, measure, analyze, improve, control. The method shows its strength in processes with many similar cases and measurable characteristics, for example in manufacturing or invoice checking. For smaller companies the full toolkit is often too big, while DMAIC as a way of thinking fits almost anywhere.
Value stream mapping
Value stream mapping shows the entire material and information flow of a product from order entry to delivery on one picture, with processing times, waiting times and inventory. It makes very clear how small the share of value-adding time in the total lead time often is. A future-state value stream follows. The method was developed in manufacturing and carries over well to order processing and administration.
Removing media breaks
A media break happens whenever information changes its medium: from paper into the computer, from an email into the ERP, from someone's head into a spreadsheet. Every change costs time and invites typing errors. Finding and closing these breaks is the most down-to-earth method of process optimization and in mid-sized companies often the most effective one. A mandatory field, an interface or a form filled in directly in the system is often all it takes.
| Method | Strength | Fits especially when |
|---|---|---|
| Lean | Makes waste visible | the workflow has many waiting times and handoffs |
| Kaizen / continuous improvement | Many small improvements in daily work | the team itself should improve things and a fixed rhythm is possible |
| Six Sigma | Reduces defects and variation measurably | there are many similar cases with data |
| Value stream mapping | Whole flow with times on one picture | material and information move across several areas |
| Removing media breaks | Quick, tangible results | data is typed in more than once today |
How to Approach Process Optimization in an SME
Companies with 20 to 250 employees rarely have a dedicated process management team. Process optimization runs alongside daily business, and the approach should fit that.
Pick a process that happens often and where a delay is felt, for example order processing, incoming invoices or complaints. Name one person who owns it and has some time to spare, and give the effort a fixed timeframe of a few weeks. Involve the people who carry out the process from the start, because they know the special cases and will carry the change in daily work. Write down the results of each phase, even if it is just one page.
Example: order processing at an engineering supplier
Illustrative example: it summarizes workflows and findings that are common in companies of this size. The company and the figures are a worked example, not a customer project.
A supplier of custom machine parts has 80 employees, six of them in inside sales. Management wants to shorten the time from purchase order to order confirmation, which often runs to four or five working days.
In phase 1 the head of inside sales talks to two colleagues in inside sales, to production planning and to purchasing. It turns out that processing an order takes about 30 minutes. The rest of the time the case waits: for technical clarification by production planning, for a delivery date from the material supplier and for approval by the sales manager, who signs every order confirmation.
Phase 2 brings three findings. Technical queries go by email to a shared inbox that is read twice a day. Purchasing asks for delivery dates on standard material every single time, although replenishment times have been stable for months. In practice, the sales manager's approval mainly checks prices that come from the price list anyway.
In phase 3 the target process takes shape: technical queries go to a named person with a deputy, standard material gets fixed replenishment times in the ERP, and the sales manager only approves orders with special prices or above a value threshold. Phase 4 starts with a pilot for one product group, and phase 5 measures the lead time again after six weeks. None of these changes needs new software, and all three fit into one process description.
Every Optimization Starts with the As-Is Process
Each of the five phases builds on the first. Optimizing without capturing the as-is process means working from an idea of how the process runs, and that idea usually comes from management's point of view. The waiting times, detours and special cases where an improvement starts are known to the people who carry out the workflow every day. Without a baseline, phase 5 also has nothing to show what has improved.
That makes the capture the part that deserves the most care, and this is where wissa.ai comes in. An AI interviews your employees in chat, in sessions of around 15 to 20 minutes, whenever it suits them. It asks about special cases, systems and handoffs and suggests inviting the colleagues who are mentioned in the conversation. The conversations turn into process documentation, an overview of all areas and processes and the findings phase 2 starts with: media breaks, duplicate work, single points of knowledge and contradictions, each finding with its source. The guide on AI process documentation describes how this works, and the page on process documentation software gives an overview of the features.
After the change, you invite people to a further interview and the documentation carries on from there. To get started there is a free trial package (10 credits, once per company, terms) that is usually enough for a first interview, and the employees you invite do not need an account. Feel free to try it with the process you want to improve next.
Next Step: AI
Artificial intelligence can speed up several of the five phases: in the capture through interviews, in the analysis by comparing many statements, and in the process itself by taking over individual steps such as document capture or sorting the inbox. The guide on AI process optimization shows which use cases have proven themselves in mid-sized companies and where the limits are.
Conclusion
Process optimization works in a clear order: capture the as-is process, analyze weak spots, design the target process, implement, measure. Lean, continuous improvement, Six Sigma and value stream mapping provide proven tools for it, and in mid-sized companies closed media breaks, clear responsibilities and sensible approval thresholds often bring most of the improvement. What carries the result is a capture that shows how work really gets done, and documentation that grows with the improved process.
FAQ
What is process optimization?
Process optimization is the planned improvement of an existing business process against goals set in advance, such as shorter lead time, fewer errors or clearer responsibilities. It starts by capturing the current state and ends with a target process that people work by and that is measured.
What are the 5 phases of process optimization?
The five phases are: capture the as-is process, analyze weak spots, design the target process, implement, and measure and adjust. Their logic matches the PDCA cycle and the DMAIC cycle from Six Sigma, and they can be repeated as a loop.
Which process optimization methods are there?
Common methods are Lean management (removing waste), Kaizen or continuous improvement (small improvements in a fixed rhythm), Six Sigma (reducing defects and variation with data) and value stream mapping (showing material and information flow with times). Simple tools such as the swimlane diagram, SIPOC, the RACI matrix and the targeted removal of media breaks complement them.
How do you approach process optimization?
Pick a frequent process where delays are felt, name one owner and set a fixed timeframe. Capture the as-is process with the people involved, collect weak spots with their source, solve organizationally first and technically second, and measure the result against the baseline.
What is the difference between process optimization and continuous improvement?
Continuous improvement is one form of process optimization: many small improvements in a fixed rhythm, usually suggested by employees themselves. Process optimization is the umbrella term and also covers one-off efforts in which a process is redesigned from the ground up.
Which metrics work for process optimization?
A few figures are enough to start: lead time, share of waiting time, first-time-right rate, error or complaint rate and the number of handoffs. Collect them before the change so that a comparison is possible.
Sources
- ISO 9001:2015, clause 4.4 (Quality management system and its processes), 9.1 (Monitoring, measurement, analysis and evaluation) and 10.3 (Continual improvement)
- Taiichi Ohno: Toyota Production System. Beyond Large-Scale Production, Productivity Press, Cambridge (MA) 1988
- Masaaki Imai: Kaizen. The Key to Japan's Competitive Success, McGraw-Hill, New York 1986
- Mike Rother, John Shook: Learning to See. Value Stream Mapping to Add Value and Eliminate Muda, Lean Enterprise Institute, Brookline (MA) 1998
- W. Edwards Deming: Out of the Crisis, MIT Press, Cambridge (MA) 1986