Next Gen AI Driven Organization — A wakeup call!
In many decisive battles, battles lines are drawn, strategies are made, and a winner emerges. However, sometimes, pitifully, battles are lost because of utter ignorance — ignorance of even where the battle front is — and more sadly, self-deception.
A brilliant piece of analysis from Forrester reports that 40 “insight-driven companies” are going to grab $1.8 trillion by 2021 — most likely a part of this is going to be carved out of market cap of your organization. In this list we have young companies that are less than 8 years old. What unifies them? Their obsession with data and AI.
Broadly, with respect to AI adoption, organizations fall into one of the two categories:
First, we have the “talkers”: there are organizations wetting their feet on what they typically call “AI initiatives” — taking small risk averse steps in organizational silos, getting tangled by bureaucracies and a minority few unfortunately focusing more on press coverage than actual outcome.
Then we have “Do-ers”: These are the insights driven companies, that have integrated (or on a strong path to integrate) Analytics & AI into their organizational fabric. These organizations have a holistic approach to what I would like to call “AI enabled Value Chain”.
Which one are you and where do you want to be?
The rest of the blog is about how to win AI the battle by becoming a “Do’er”
1. Aim for an AI driven value chain
Three stories about visionaries who embody the idea of deeper and broader AI adoption to drive true innovation across their organization.
Enter Stitch Fix. Starting an online cloth retail business in the midst of giants like Amazon? No problem — founded in 2011, now stitch fix is valued at $4.4B. How? Conscious focus on data & AI.
All retailers get few standard data points from brand manufacturers: size etc. However Stitch Fix invests effort in extracting upto 80 attributes per clothing item like collar type, number of buttons etc. They use Machine learning to estimate photo based detailed clothing measurements (e.g. sleeves and pant leg profiles). The data points give them a competitive advantage when they match (recommend) clothes to people (they have around 75 data points per person). You have a penchant for double collar shirts or wooden buttons? No problem — they can figure it out.
They apply data and AI to entire value chain: from warehouse assignments, matching customers with human stylists and upto the level of creating designs for new clothing items.
A posterboy in this area is Netflix. The movie recommendation-engine is only one part of the story. Consider this: Industry success rate for new tv shows are 35 % — however Netflix has 70%. How? Before green lighting the first biggest production of House of Cards ($100m investment!), Netflix turned to its past data to predict the success rate and factors. The rest is history. Netflix even optimizes on the posters that will be displayed when you browse, it uses AI for quality control (identifying low quality media) and for optimized content delivery (which media to ship to your nearest Content Delivery Network)
The last story is about Earnest. A Digital Lender Launched in 2014. Their business is in giving loans at lower interest rate than the traditional financial institutions that has been around for a time. How? By risk more holistically. Their algorithm considers around 100,000 data points about the applicants — including non-traditional data points like education, career trajectory etc. As of Oct 2017, it has extended $2B in student loans and had roughly a $500M loan portfolio.
Takeaway: Think big, think strategic –ask how can data & AI help you innovate in the various aspects of your value chain. Of course when you execute, you will prioritize — however the key point here is to have the big picture in front of you, so you can effectively percolate the vision throughout the organization and empower the execution of the vision.
2. Executives Sponsors who understand the possibilities beyond the hype
AI is undoubtedly hyped. In the words of a prominent leader in this area, Michael Jordan (Professor at University of California, Berkley): “Most of what is being called AI today, particularly in the public sphere, is what has been called Machine Learning (ML) for the past several decades”.
From Judea Pearl who won the Turing Award, the highest honor in computer science: “All the impressive achievements of deep learning amount to just curve fitting”.
On a lighter note, I read this true but funny quip online: Few years ago people talked about “training their ML models” — now people are talking about “teaching the AI system to understand images”.
However beyond this hype, there is a LOT of possibility with the current technology today: help identify tumors faster, semi-autonomous cars, personalized marketing, identifying fraud — I can keep going. Wherever there is historical data to learn from, we can potentially benefit by leveraging Machine Learning.
I believe the Executive Sponsors should understand these boundaries, so they can steer clear of the hype and deliver tangible business value by exploiting the area of the possible. Enter Katrina Lake — Founder and CEO of Stich fix. Katrina was trying to woo Eric Colson, who was VP of DataScience at Netflix to join Stich Fix. After few meetings, Colson concluded that “the person across the table was an intellectual clone of his longtime boss, Netflix founder Reed Hastings”. Colson says “You feed them a little bit of information and they can paint a vibrant picture that matches reality”. (source: inc.com)
Takeaway: Execs take note: spend time with experts in this area to understand the fundamentals of AI. This will help you navigate around the hype and deliver tangible value.
3. Think holistically about AI: Risk & Responsibility
Adopting AI & ML in your organization is much more than optimizing for a metric (clicks, views, sales etc). Yes, Business Value is one important dimension in AI adoption — however there are other key areas that needs be taken into consideration: manage risks (for your organization) and responsibility (to the society).
A story about Risk:
18-March-2018, 9:39 pm, 15C temperature: An Uber Self driving car with a driver inside kicks into autonomous driving mode in Tempe, Arizona. It will be in autonomous mode for the next 19 minutes.
9:58 pm: 49 year old Elaine Herzberg pushing her bicycle crossing lanes. The car crashes into Elaine killing her.
Quote from SFGate: “There is no question the laser should have seen her,” said Brad Templeton, a Silicon Valley entrepreneur who was an early consultant on Google’s self-driving project. “I know the technology is better than that, so I do feel that it must be Uber’s failure.”
Outcome: One life lost. Uber pulled down its autonomous driving fleet for a while to make things safer (and lot of bad press).
Question: If there were robust training data & comprehensive testing process (under different light/weather conditions), could this been prevented? Should we had technology to monitor the Safety Driver? i.e. is it worthwhile asking lots of questions before letting this run on the road?
A story about Responsibility
People are googled all the time: before a meeting, for job applications etc. A colleague of Latanya Sweeney googled her and found and ad stating “Latanya Sweeney, Arrested?”
Latanya Sweeney being a Professor at Harvard & known for uncovering challenging issues in the past, researched and found statistically significant discrimination in the ad delivery systems based on searches of racially associated personal names. She went ahead and published a well-regarded paper.
Takeaway: As a part of the AI strategy, organizations need to consider both Risk (Governance, Risk & Compliance) and Responsibly (Fairness, Privacy etc). If this is not done voluntarily and upfront, there will come a time this has to be done under pressure and from a very bad spot. I will be sharing more inputs in this area in my upcoming blogs.
Parting words
In this blog I have outlined three aspects that I believe are essential in getting your share of the $1.8T by 2021: Aim for AI driven Value Chain, Execs Sponsors need to focus on the possibilities beyond the hype, and design a holistic AI Strategy — encompassing Risk & Responsibility in addition to Business Value
The question is: are you going to get your share or are you going to be carved out? A quick press release makes you happy or are you willing to invest in holistic approach to AI Strategy? Are you ready for the long haul? Are you a Doer?
