Creating an AI Roadmap

In the famous words of Peter Drucker businesses need to "innovate or die".

Today that means automation and Artificial Intelligence.

We helped PCCW do just that through our unique AI Innovation Workshop we worked with the leadership team to create their own roadmap to Artificial Intelligence.

The challenge

PCCW is Hong Kong's leading premier telecommunications company with a vast number of mobile customers. In such a competitive market place, customer retention is a major battle. We worked with the Customer Services leadership team to look at how AI could be used to solve this major problem.

Our innovation workshops are more than just a day brainstorming. We like to get to know our clients problem from the position of those dealing with it day to day. We surveyed the staff to learn more about how they work, and what their biggest issues are.

Once the results where in we had a basis to start our workshop form. The day began with a brief introduction to Artificial Intelligence. Based on use cases that competitors or parallel markets are developing AI for. This helped to frame the rest of the days workshop.

Next a deep dive into the problems by analysing the questionnaires and holding a round table discussion with the team.

The problem was framed. PCCW need to know more about their customers to be able to predict why they churn. This problem is exacerbated with disparate systems and an overwhelming amount of user data. But not necessarily the correct or insightful data, they didn't know.

The end of the day, we talked about potential AI solutions that could be applied, and what the path to get there was. Discussing areas of data access, cleanliness and management, and how important that is as a building block to AI.


The result

The rest of the week was spent reviewing data processes and capture. Where we then put together a recommended AI roadmap. This outlines the now, near and future goals for the team to work on. Including getting their data aggregated into a data lake to ensure a data scientist can carry out effective modelling (sequential modelling in this case). Then identifying the types of data that would be useful for a churn prediction model, finally how that could be built, tested and deployed.

“Skim Technologies came with fresh ideas for our team. We didn't know that AI could be used to solve some of these problems we have. But it was very easy to understand how we can now build and use AI in our business.”

Kurt Chi, Customer Services Director, PCCW

Our mission

Skim’s mission is to empower people to use data more effectively and to demystify artificial intelligence. Rather than holding up the common narrative of machines replacing humans, we see how machines can help humans to have easier lives and better businesses.

Supported by


London office
27 Finsbury Circus,
London EC2M 5NT

Portugal office

+44 207 129 7497