For process practitioners

You know what a process is, you have seen process maps, and you may well have used a commercial process intelligence suite. You do not need the concepts. What you need is to get from “we suspect something is wrong in order-to-cash” to a finding somebody will act on — this week, on a laptop, without a project.

What is different here

Coming from a hosted platform, three things will feel unfamiliar, and all three are deliberate:

• There is no data upload. The log is parsed and queried on your own machine. For a first analysis on production data this usually removes the longest step in the whole project, which is getting permission.
• There is no connector catalog. You bring an export — CSV or XES. That sounds like a step backwards until you notice it means you can start today with a query someone in IT runs once.
• There is no dashboard to configure. Analysis produces artifacts you keep, compare and hand over, not tiles you arrange.

It scales, but start small

Start with an export of a few hundred thousand events on your laptop. Everything in this track works identically on a much larger log — the difference is where it executes, not how you work.

The track

1. Scoping a first analysis — choosing a process and a question narrow enough to answer.

2. Getting event data out of a system — what to ask IT for, in one paragraph they can act on.

3. From process map to a decision — reading a discovered model without over-reading it.

4. Rework, waiting time and conformance — the four findings that pay for the exercise.

5. Sharing results with the business — making a finding survive contact with the people who own the process.

How long this takes

A realistic first analysis, with data already in hand, is a day or two. Getting the export is usually the long pole — often a week or two of other people’s time, which is why the extraction page is written as something you can forward.