AI/ML: A Starter Guide for Business

Let me start my second blog for AI for business. If you have n't read my first blog - Data Science in Reality - Please read it - Here. I have talked about the Data Science Focus points, production tidbits, Framework, and skills required to be efficient.  

Just some ground rules - I am using the term AI/ML, and if you are reading any other blogs, different people are using AI. In my opinion, AI is the buzz word may help to attract stakeholders or investors eye-balls. But, in reality, we are not able to do AI in my life-time. So, people using the word AI term every-time, think twice!

Put some definition for the reader:

·    Artificial Intelligence -- A category of technological pursuits that hopes to replicate how humans think -- that is, what makes them intelligent.

·    Machine Learning -- A branch of probabilistic computational techniques that have yielded much of the recent progress towards Artificial Intelligence in the last decade. A field of computation that uses algorithms to derive probabilistic models of complex events by training these models with a set of observations (data).

Mindset Shift

As I have talked with the business decision manager or leaders, one of the typical feeling/sentences I get is :

"ML is hard..."

"ML doesn't fit our priority.."

" We are using the same process; why should we change it..!?"

Well, I think its a question of changing the mindset and explain what does is AI/ML.

• Traditional Programs - a set of programs and focus on the rule-based program. Emphasis is on the calculation.
• AI/ML - Based on data and focus on Intuition. Intuition is acted based on experience and experience based on historical data input, action, feedback, change behavior, adapt, and a lot of loops.

We can leverage both by combing calculation and Intuition in the real-world. It can be applied not only to software / Technology companies but from healthcare, retail, manufacturing and others to helps in automate, understand, build, relate, and onboard.

Show the AI/ML Value

Now you may have convinced your stakeholder. Show the value of AI/ML value by running an MVP (Minimal Viable Product). We focus on building something that is not only working, but it should be integrated and scalable in business processes or products. 

A couple of points:  

• Have a clearly defined and measurable objective that creates business value. 
• Partner with an internal team or build in-depth domain knowledge in the business and show some tractions by building the MVP.
• If you have multiple ideas for the business - go for low hanging fruits first and measure the business value by creating a storyboard. Business Value KPI can be HEART ( I will come back to this in PM Metrics Session). 
• It is an iterative process - build trust with the business partner, show the product, take buy-in, move. ( It is hard iteration, trust me!)

Building AI/ML Culture & Team 

In my earlier post, I mentioned you need to have all sorts of skills to be productive here. Again, you don't need to be an expert.

Build trust, build a relationship, build trust

AI/ML is a multi-discipline team that needs to work together. So, you need to have enough knowledge to work with the teams. As you are building the AI/ML team - You need to be excellent in the data domain, but you need to develop the data culture around yourself. Work with stakeholders to teach a tidbit of how AI/ML can help bring the business value.

Forming the Core team

When you are building the core AI/ML team - I would suggest creating an in-house team. The internal team has business knowledge and domain expertise. It scales very fast with minimum effort. Look inside your organization, and there is the talent - may just need grooming :)

AI/ML Training

Well, I divide this section into three bullet-points:

Business & Executive leaders :

If you're an executive - I will urge you to take learn what AI/ML can do for the enterprise in terms of Framework, Strategy, and Value. 

AI Leader, AI project/product Manager:

Project Leader will able to set the AI in the project, team formation, translate business problem to AI/ML problem, Story mapping for various business problems and decide based on the business value, architecture and workflow for scalable and integration. 

Data Scientist / AI / ML Engineer:

• Deep understanding of Business, technical knowledge of Machine Learning
• Open source tools and libraries are excellent. But, understanding the math's is also necessary. (Yes, I am all for Business - not research. But, but, you need to understand the math's behind it. I will give you an analogy of car/bike - If something breaks, you need to know what is wrong with it and take action - data, algorithm, etc.) 
• Hands-on expertise with team, integration, work, deployment and see the business value
• On-going education with the recent up-to-date advancements

Build AI/ML Framework & Strategy  

Framework >> Strategy >> Business Objectives

When you are thinking of building the AI/ML framework, you need to think of multiple dimensions, like:

• Platform : What kind of platform you would like to make: In-premise or cloud provider. What should be the right provider based on price, time, correct skillset, enhance competitor advantage?
• Tools : Which tools are right for the solution provided by a cloud vendor, open sources? Is it aligned with the current technology stack?
• Business process : Start business process which process needs to automate? Business process flow/architecture.
• Team structure : In-house team - Team alignment & structure.
• Business capability : Enhance building capability - core vs. non-core.
• Organization structure : How to align the AI/ML team with each of the business units?
• Internal /External partner : Build the relationship, interactions with the eternal team. Do you need to build new relationships or vendors for AI/ML?

As a leader, you can align with business/company OGSM or OKRS and build a strategy around each dimension. Again, the above list is not comprehensive. But I hope you got the point!

The Framework gives guidance, and sets of the decision you make are AI Strategy.

Framework + Decision Making = Strategy

As I am closing this article - Here are the top 6 things to remember:

• DATA - Bread, butter, Oil: Asset for the organization!
• Business / Domain Knowledge is the real deal.
• IT system - Think of people, governance, and culture challenges.
• Enterprise architecture transformation
• Change of mindset: based on the Intuition (probabilistic approach)
• Leadership drive - the top-down approach is required for a thriving AI/ML culture.

In my next article - I will dig deep into one of the skill sets from previous post ; stay tuned!