Process Optimization Management:
Part 5.2 — Improve Phase / A.I. Scoping Workshop

Picture from Patricia Alexandre at Pixabay

After we defined artificial intelligence in the last chapter, we will turn today towards our A.I. scoping workshop.

Excluding Non-A.I. solutions

Imagine you and your team are sitting together with bright data scientists discussion how you can use A.I. to solve your previously defined problem.

The first step in this scoping workshop is actually to exclude all aspects which are not A.I. solutions. Many people find that counterintuitive, but reality shows that many problems can actually be solved by not using A.I. first.

Let me dive into this a little bit more. Let’s consider the topic of our two baristas, Betty and Michele, which we talked about in one of the previous articles. Betty is much better in working the machine than Michele. We spotted that in our analysis phase and can now try to solve it with A.I. (write algorithm to implement it in the machine to counteract “bad” handling) or we could just have Betty train Michele. If Michele is untrainable, we will offer a job as a cashier or let Michele unfortunately go. As a rule of thumb:

if you can solve it without A.I., solve it without A.I..
Reduce complexity in problem solving and try the easiest way possible.

I need to stress this, because people in general will suggest solutions close to their own profession. If you ask an engineer, he will give a solution using engineering. If you ask a data scientist, she will give a solution using data science. If you ask a business manager, it will give a busines solution. That is one of the reasons, why we at Goldblum Consulting tend to support our clients in these scoping workshops to find the best solution. And by best I mean best for your company and yourself.

There are several levels in problem solving, we will look at during the A.I. scoping workshop:

- Solving problem by changing business processes (e.g. change supplier or training staff)

- Using robotic desktop automation

- Using robotic process automation

- Using Business Intelligence (descriptive statistics) — visualising data

- Using Business Intelligence (inferential statistics) — inferring and predicting

- Using Artificial Intelligence

We will talk about each of the above topics in more detail over the next weeks, but for now we will just name them, so that you have an understanding what “non-A.I. solutions” could be.

Carving out true A.I. based solutions

Good outsourcing partners will be supportive in separating the non-A.I. from the the A.I. solutions. Follow your guts in these workshops. If they start feeling like a sales pitch, just leave. But if you feel understood and supported, stay .

Once you found true A.I .applications, your outsourcing partner will be able to play out his A.I. muscles. First of all, you will determine to what cluster of A.I. problems, your problem belongs. Will it be for example a sensing problem like in image recognition or a prediction problem in finding for example potential fraud cases. Or is it a forecasting problem or a case of processing of unstructured data? Or is it NLP (natural language processing) where you try to process emails, SMS or PDFs?

Each cluster of A.I. problems has its standard set of algorithms and specialists. By clustering your problem first, your outsourcing partner will be able to assign cluster specialists to your problem and might be able to use standard tools and algorithms to tackle your problem. In 2021 we are far enough into A.I. that the wheel does not have to be invented with every project you are doing. Using standard tools will save you time and money.

At the end of the workshop, you will have a good idea, how A.I. will be able to help you and you will have created a rough timeline for a lean prototype to be developed. Furthermore your A.I. partner should have a good understanding where your A.I. solution will “live” in your company and who should use it. They should have a very good understanding, how your employees will need to work with the A.I. solution and in what way it will be implemented in your company.

Tomorrow we will look at the last phase — the control phase. This phase will insure that your amazing A.I. solutions will continue to live in your company and not just collect dust in the virtual software shelf.

If you are interested in A.I. leadership education or want to start your A.I. journey, just contact me at ansgar_linkedin@goldblum-consulting.com

INDEX

  1. Overview
    1. Process Optimisation Overview

5-Phases of Process Optimisation

2. Define Phase
2.1 Define Phase — Part 1
2.2 Define Phase — Part 2

3. Measure Phase
3.1 Measure Phase — Part 1
3.2 Measure Phase — Part 2

4. Analyse Phase
4.1 Analysis Phase
4.2 Analysis Phase

5. Improve Phase
5.1 Improve Phase / Defining A.I.
5.2 Improve Phase / A.I. Scoping Workshop

6. Control Phase
Control Phase

AI Evangelist — CEO of Goldblum Consulting