TL;DR:
- Not all businesses require an AI Strategist; an established data management process and optimised reporting are prerequisites.
- AI is just one tool in the data toolkit, and not all business problems require AI solutions.
- A Data Strategist with expertise in data management, analytics, and machine learning might be a better fit for most companies.
Thoughts from Data Scientist expert, Koo Ping Shung
Hi, I’m Koo. I’ve been working with Data and Artificial Intelligence for over 20 years. I’ve done all sorts of things like collecting data, managing it, and making sure it’s used properly. I also help to find useful information from data and put it into machine learning models. Every now and then, I notice what’s happening and what’s difficult in this field and I write about it.
Lately, I’ve seen a lot of people using the title “AI Strategist”. I also get asked about this a lot at my company Data Science Rex (DSR). So I thought it was time to share my thoughts.
So… grab a coffee and settle down as we unpack whether your business really needs an ‘AI Strategist’. Here’s a humble view from a Data Scientist:
The Prerequisites of AI Adoption
Firstly, there are certain conditions that need to be met in your business before you start thinking of hiring an AI strategist.
1. Data
We all know that AI works better when there are good quality data, followed by availability of relevant data. So, you need to assess whether your business has an established data management process first. What is an established data management process then? It should have the following:
Infrastructure:
Are there established data storage and retrieval infrastructure? It needs to be something like a data warehouse, a SINGLE place where all the datasets go to be stored and managed.
Data Management Processes:
Are there data management process that accompanies the infrastructure? In order to ensure your data is of good quality, it needs to be managed well. This is not forgetting that there should be processes in place to ensure high security and privacy level. These are processes that deals with Identity & Access Management or IAM in short.
Data Quality Measurement:
Are you measuring data quality? Data quality needs to be measured to build up confidence in using data. Having established data quality metrics helps in showcasing to your stakeholders that the reports or derivative products from data such as insights, visualisations and analysis can be trusted.
2. Reporting
Have your organization optimised your reporting process using data? Why is business reporting important before coming to AI?
We need to actively understand the nuances of data by examining how and when the data is collected and determining what each column in the dataset represents.
Only through a good understanding of these nuances, that are honed through multiple reporting, can we have a good understanding on what are the limitations from internal data.
Establishing AI through your business requires change management extensively. This means that throughout your business, there is a need for everyone to understand how data and AI works, followed by everyone feels that they have benefitted from using data.
An optimised reporting process throughout the company is a good signpost to say that your business and its employees have benefitted from data for their own tasks.
Once these conditions have been met, there’s a good chance that if there are AI use cases in your organisation, you may be successful. The reasons are it has tackled two of the biggest roadblock to adopting Artificial Intelligence in business, Data and Change Management.
AI is just part of the data toolkit. You can see AI as, besides the ubiquitous tools that you find in the toolkit like hammer, screwdrivers, etc, that smart shiny drill that can auto-change the drill bit based on your needs.
From this analogy, you will realise that not all your business challenges are about “drilling” i.e. AI solves certain type of problems, not all problems. What does this translate? Not all the business problems you have needs AI, and followed by do you have enough AI use cases to justify hiring a full-time AI strategist?
Do not forget that AI use cases like any IT projects has costs to it, short-term being the proof-of-concept, design & planning, data management, training of the AI models, and long-term being the maintenance, constant monitoring and validation of AI models, cloud computing subscription, ensuring the essential skills stays within the organization.
Switching Perspectives
Now put yourself in the shoes of an AI strategist, let us forget about his/her background, experience and suitability for the role for the moment.
The AI Strategist main role is to identify areas in the business where AI can be used, proposed the necessary changes that needs to be made, and how AI should be integrated into the business process, keeping in mind the alignment to strategic goals of the business.
Have you ever heard of the saying, “To a hammer, everything is a nail.” To an AI strategist…everything must use AI? Wait a minute! You probably realise right now, a few paragraphs before, we did mention that not all use cases need AI! AI is just one of the tools in the data toolkit! If a simple average is needed to solve a business challenge, why go through all the hassle and the costs (short- & long-term)?
4. So What Do I Need?
Data will be the new normal, and taking advantage of data will be what good business leaders will constantly be thinking about. There are multiple tools, with variation in value derived from data, that business can take advantage of, and Artificial Intelligence is part of it.
What you really need is a Data Strategist, that can help your business see end-to-end from data collection all the way to change management when the derivatives of data, such as decision models or insights, are being used in business processes.
I could write a whole article on how you should select a Data Strategist for your organisation (let me know in the comments below if you’d like to read this), but to give some quick pointers, the person needs to have background in the following:
- Data Management & Data Quality
- Data Analytics & Data Analysis
- Machine Learning, Model Deployment & Model Monitoring
- Generic Business Processes & Change Management
I can tell you very quickly such talents are not easy to find. But a word of caution, do not jump onto one because of the titles. Titles are easy to award, but of most importance is their experience and background! Be comfortable with it.
And please remember:
“AI is not a one-size-fits-all solution, and other data tools might be more suitable for specific problems. It works better when there are good quality data, followed by the availability of relevant data.”
I wish your company all the best in building up data capabilities and taking advantage of data analytics and artificial intelligence (if any). –– Koo
Have you considered the benefits of a Data Strategist for your business instead of an AI Strategist? Share your thoughts below and don’t forget to subscribe for updates on AI and AGI developments. You can also connect with Koo Ping Shung on LinkedIn by tapping here.
You may also like:
