Making progress
a practice.

When someone raises a problem, the leader’s response matters. Asking what happened and taking time to understand the work makes it easier for people to speak up next time. Jumping straight to an answer can have the opposite effect, even when the intention is to help.

I see coaching as part of the daily job. A team needs to know what is expected, how it is performing and which problems need attention. The leader’s role is to ask questions, help remove obstacles and follow up. Solving every problem for the team leaves people dependent on the same few individuals.

Daily management gives those conversations a regular place. It should help people understand the process and decide what to do next. If a meeting consists mainly of reporting numbers upwards, it is worth asking what the team gets from it.

Clear expectations and respect for people belong together. We need to be able to discuss a missed target without making it personal. We also need to check that people have the training, time and authority to do something about it.

I pay attention to follow-up. Returning to an issue, asking what was tried and listening to what the team learned shows that the work matters. Those conversations are easy to postpone when the schedule gets busy. Making time for them is part of leading improvement.

A Kaizen event gives a team time to look closely at a problem. People from different functions can compare what they see, try changes and get decisions made. That focus is useful, particularly when the same issue has been passed between departments for a long time.

The difficult part often comes after the event. Production pressure returns, people change shifts and the team has other problems to deal with. A change that worked during a focused week may need adjustment before it works reliably in normal conditions.

I want follow-up to include the process itself. Are people using the new method? Does it help them? What happens when volume changes or an experienced colleague is absent? Closing an action in a spreadsheet tells us less than watching the work and asking the people doing it.

Standard work helps when it describes a method people can actually use. It also needs an owner and a way to raise problems. Otherwise a document can remain technically current while the work has already moved on.

Daily management is where I would expect the team to discuss setbacks and agree on support. Sometimes training is needed. Sometimes equipment or another process is limiting the change. A useful follow-up makes that visible and assigns responsibility. The event creates momentum; the following weeks determine whether the improvement holds.

I start by looking at the task. Where are people spending time collecting information, searching documents or preparing a summary? What do they need that is currently difficult to find? That gives an AI experiment a purpose and something to compare against.

A small first application is easier to assess. Someone should own the process, review the output and decide what a useful result looks like. I would compare the time spent, the quality of the information and the amount of checking required. A faster first draft can still create extra work later.

The person using the tool needs to know its limits. AI can produce a confident answer that is wrong or leave out context that changes the meaning. In manufacturing, those details matter. Information access, confidentiality and responsibility for decisions need to be clear before a tool becomes part of the workflow.

Lean still applies. We should understand why a task exists before automating it. Cleaning up the source information or removing an unnecessary handover may solve more of the problem than adding another tool.

I am interested in applications that give people more time to understand a problem and work with colleagues. Getting there means testing under normal conditions, listening to users and being willing to change or stop an experiment. The evidence should come from the work, rather than the demonstration.