Building Data Science & AI Capabilities in Businesses

by Koo Ping Shung

I've been asked many times how will I go about to build capabilities in businesses, capabilities that can tap onto the data collected after digital transformation.

The biggest issue right now, in most corporations is the lack of use cases, IMO. The lack of use cases, will lead to disappointment as corporations forked out massive amount of cash to hire green "talents" (that lacks the capability to determine and scoping use cases, see my previous post here), purchase software and hardware that has ended up as white elephants.

Thus my thought process here is how do we seek out relevant use cases, and build capabilities accordingly.

Step 1: Boosting Data Literacy

The first step is to make the whole organization more data-driven, starting with building up the culture. Encourage employees that uses data for their decision making. Start sharing sessions on successful presentation, that uses data to shape decision making. With these successful use cases, it will motivate other staff to start using data. It will then be appropriate to start providing data literacy courses, as staff are more motivated to learn how data can help with their decision making and career. This will cast a wider net to reel in more data use cases, with increased understanding on the strengths and weaknesses on using data.

Step 2: Collecting use cases and data

Ideally, there should be someone very experienced working with data at this step. With more use cases coming in, the experienced data professional (which can be an external consultant) to start assessing and prioritizing these use cases. Why someone experienced here is, he/she can assessed if the use cases can be done, how to collect the relevant data, etc i.e. assessed the data-problem fit, and act accordingly.

With the use cases prioritised, it is time to think about the data. It takes time to collect good quality data, and time cannot be rushed, not a challenge that can be overcome by throwing money at it. Collecting good quality data need planning. Good planning will help in reducing time wastage.

Step 3: Imparting skills and knowledge

With the use cases prioritised, we can start designing and executing the training roadmap and/or hiring. The training roadmap will help businesses to focus on training talents with the relevant skills to tackle the use cases immediately. This is a more focused and customised approach as compared to sending talents to generalised courses out there. This can be an expensive approach but the use cases will provide good justification to take this approach.

Step 4: Enabling Knowledge Transfer

All projects/use cases will have their nuances. A group of "green" data scientists or data analysts might not know how to tackle them. This is where mentoring can come into the picture. Companies should get someone with experienced to mentor the group, enabling knowledge transfer, helping the data scientists to see the possible best-practices and approaches to designing solutions. This will also complete the final step in enabling the "green" data scientist to apply what they have learned in Step 3, and see for themselves the strengths and weaknesses of each tools used.

Step 5: Building that virtuous cycle

As business gains more applications and knowledge in using data, new use cases WILL BE discovered. This is where we can move to Step 2 to Step 4 again. :)

Disclaimer: I do provide such services in my business such as:

  • Identifying Use Cases
  • Data Inventory Consultation - Relevance of existing data, collection of quality data, measuring data quality, setting up data governance process, etc.
  • Training services
  • Mentoring service - Help data science teams to solve problems with data

Conclusion

If you have read all my posts, you will have understand that my biggest focus is VALUE. The value created must be greater than the costs that goes into it, cost being effort, time and money. As such, the whole design of how to build the capabilities, as can be seen, is very much focus on realising the value AND ensuring that the costs of running the projects will be justified by the value created.

What are your thoughts on this? If you are keen to reach out to me, to discuss on how your department or business organization can build up strong and lasting Data Science and AI capabilities, you can PM me on my LinkedIn.

Do consider signing up for my newsletter too. I just started my YouTube channel, do consider subscribing to it to give support! :)

Consider supporting my work by buying me a "coffee" here. :)