Source: Google Images

Artificial Intelligence for C-Suite and Business Leaders

Srivatsan Srinivasan
Sep 4 · 7 min read

Last week I happen to meet one of my ex-customers and as with any 2 guys in data space, our conversation leads to AI. Long story short, his question to me was

  • How can CDO, CAO, CIO or Business Leader understand AI/ML in simple terms?
  • How can Business Leader identify and qualify the ML problem?
  • While they have a competent team running data and AI functions, How can they talk in their language and understand their terminology?

I have added on additional roles in C-Suite to the list who needs to be AI aware. You might have guessed it right, CEO and CFO. Now my ex-customer was doubtful on the need for CEO/CFO to be AI aware but below was my justification to it

Typically it is easy for CEO or CFO to pass on AI initiatives to Chief Data or Information or Digital officer. But here are few compelling reasons for CEO/CFO to be AI aware

  • CEO is closer to changing business and competitor landscape and understanding AI applicability can help them remodel their business priorities
  • It can help them prioritize investment in right use cases and put AI at the core of their business
  • CDO/CIO/CAO (CDAIO) in most cases think from a technology perspective and might not be the best to detect business trends. Also, CDAIO need to be pushed to look at AI initiatives from an ROI perspective
  • There is a simpler and effective solution to a very complex problem. The curiosity of Do It Our-self of some CDAIO might delay insights. Will cover more in this later in the article
  • Final and most important one AI can shape organization P&L by scaling business process, optimizing operation cost, increasing revenue contribution and monetizing data assets. No one better than CEO/CFO can drive these initiatives to success

The premise of this article is to provide resources that can help C-Suite (CEO, CFO, CDAIO) and Business Leaders (Marketing, Risk, Fraud, Supply Chain, etc) get an idea on various terms and buzz words in AI space and help build thinking around identifying and qualifying use case

Let us get started. The content highlighted here will be a good 6 to 7 hours of reading or watching without going into technical details at that same time get an idea of what’s technically possible

The information I am sharing below are completely free but quality 
resources that will help get intution on AI, application of AI and on how to frame an AI problem

Below is an amazing machine learning 101 presentation by Jason Mayes. This rich in graphics visualization covers

  • What is ML and types of ML?
  • How does it work?
  • How it can be used?

Once you have gone through the presentation above (don’t miss out the link in the notes section of presentation if you want to dive into detail), check out 2 modules highlighted below from Amazon Machine Learning for Decision Makers module

Some of the content will be repeated from the ones in earlier presentation and it is completely fine. Just watch it again. I have highlighted the 2 modules that you need to focus on, skip the remaining 2

Once you understand AI/ML, it’s applicability and how to identify AI problem. Look at Microsoft module on building an AI-ready culture within your business

The final one, Just glance through Google ML crash course on problem framing. Not required to spend too much time. This is just to give intuition on how google identifies good problems for AI to solve

That’s it from a resources perspective. I am now going to focus on my earlier conversation on “Simpler and Effective solution to very complex problems”. Understand this is very important before pouring in money into AI projects

Let me walk you through this with examples. You have huge Call Center operation and it is incurring pretty high cost. You want to cut down call volume or reduce call duration to reduce the cost of operations

The first thing to get to your objective is to find out why customers are calling you and you also want to understand where they are spending more time during the call. Is it hold time, Agent inability to solve it quickly or the complex process of information retrieval

You realize that you do not have detailed notes captured for each call and you want to convert stored voice calls to text. Your technology team comes and tell you we will use open-source software to convert voice to text

One thing AI leader needs to understand is the complexity to get an accurate text out of voice is tedious and requires a long cycle of training and tuning open-source framework. Even if open source software is trained you might not get a high level of accuracy. This is one example of a complex problem that is already solved by leading cloud providers and they expose their services as API that can be used based on pay as you go model.

This allows the organization to accelerate AI initiatives at a faster pace. There are similar API’s and services present for Natural Language processing/understanding, Images, translation, maps, OCR, etc.

Understanding these services allows one to take a call on what approach is right for AI initiative considering a priority, timeline, ROI and long term vision into consideration

Another example is in insurance where you might have to extract details out of documents to settle claims. Now custom build using open source is not going to be as accurate in some cases as some cloud API’s. The ongoing implementation and maintenance cost is also going to be high for a custom build. But with cloud API’s even if you have a million documents to process per month it will cost you hardly $1500 per month

It is very important for a technology team to look at simpler, cost-effective and yet reliable solution in the process of implementing a use case

These API’s might not be perfect in all cases. Try it out quickly for your use case and decide on buying v/s building.

The final one I am going to talk about is how C-Suite and Business leader can Identify and Prioritize use case that generates business value in longer term.

It is difficult to size and budget AI projects due to its uncertainty on the outcome, possible longer time to experiment than actually accounted, complexity in integration with business process, training deployment skew among others. This makes AI project difficult to budget and to measure return on investment

To address the above challenge I have created “AI Value Realization Framework” to calculate ROI on AI projects. The details on this framework can be found in below link and also the template is available in open domain for anyone and everyone to use it

Part 1 of the article setting context on ROI calculation and link to template is below. Download the ROI template customize it to your organization needs

I have put some optional reads in appendix section below mostly to understand why AI projects fail and importance of data strategy for successful AI strategy

If you are looking for information on how to set up and build AI organization, connect or follow me on LinkedIn (https://www.linkedin.com/in/srivatsan-srinivasan-b8131b/) to get notified on my future articles

Appendix:

https://www.linkedin.com/pulse/think-data-first-before-being-ai-srivatsan-srinivasan/

Data Driven Investor

from confusion to clarity, not insanity

Srivatsan Srinivasan

Written by

Data Scientist | Data Engineer

Data Driven Investor

from confusion to clarity, not insanity

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade