I'm excited to share my journey with you, involving the brilliant minds I encounter and the emerging issues that need tackling. As the AI community grows, it's important to address some of the gaps we encounter in this nascent and highly experimental phase as companies, organizations and countries continue to figure their place in the world of AI.
The last decade of my career pushed me in the direction of social media and big data. When I launched Yahoo! Answers in Canada that was a pivotal time for me – a defining moment that turned everything I knew about marketing on its head. The essence of community became known to me and the dynamics that cultivated these relationships before my eyes was insightful as well as fascinating. For once in my life I realized that my view of marketing as a catalyst for behavioral change was dismissed. Within these percolating community discussions, what arose was my personal window into the lives of people who put themselves out there, online, with raw authenticity and undeniable trust in the community that was organically cultivated into "family." The influence of the "establishment," while perceived to be the main driver of information and impact, could not compete with this rustling underground of discussion and diverse opinion that trickled below the fold. They dismissed the voices and opinions of those who controlled the airwaves for far too long.
My grounding is in database marketing. Data was a driver of everything I did. My first jobs in direct marketing, big agency and banking, all leveraged the knowledge of the data to drive decisions. I dug in and became pretty adept at analyzing customer behavior and determining attributions that correlated with business outcomes: in customer churn, engagement and conversion. It was rudimentary work but it did the job. I worked in the credit card industry for five years. We knew everything about our customers. Transactions, when pieced together nicely, tell a story about a person's tendencies, the things they value, their in-the-moment instinctual habits, and how responsible they are with their money. At the time, the latter was the most important thing for the bank. We just wanted to sell more credit cards with measured risk. The analyses were onerous and the implications and conclusions were sound – or so I thought. The reality was the information I used to analyze was based on decades of best practice that defined business profitability. It was also finite because not only did I not have access to a full understanding of how a customer "got here," I was handcuffed with analyzing ONLY the LOB (line of business) data. At the time I didn't question it. But soon I grew skeptical of how to effectively target customers with the right offer if I didn't have ancillary information about their "other" accounts, the other debt, the life events, influences and the customer receptivity to my offer.
I finally left banking and the bricks and mortars business to dive into the online world at Yahoo! From that day forward I experienced a world of constant experimentation, the desire to try new things and fail, and then try again. Start-up life meant that we could also question the things that we came to know and hold dear... and unlearn them in the process. My journey into social media and big data transformed the way I thought about industry, its limits and lack of knowledge about its customers, and its arrogant belief that doing "what has always been done" will allow them to sustain the relationship with customers and acquire new ones. Not so. Unfortunately, these practices continue today. I am not surprised that companies still employ a spray and play approach to ad targeting. The goal is ROI and that is driving the short-term tactics. The C-suite doesn't really understand the customer because it takes too long and it falls outside of the parameters of how performance is measured. The thinking goes, "If i acquire X number of customers at a cost per acquisition of Y" then I would have achieved my goal. What no one understands are the long-term implications of short-sighted decisions.
The world has figured out quickly that the voices of the people are more powerful and drive significant influence to disrupt business. The problem is that "legacy" is hard to change. Archaic process, job descriptions, and even culture are impediments to business keeping up with the market demands.
Where I've evolved in the last 5 years is to a place to really dive deep into understanding who we are as consumers and people. Customer Journey is a buzz word and has been for some time. I would argue that it's still a mythical concept, in part, because no one company is doing it right. I envisioned a time when we could marry social data with transactional information to get a true sense and empathy for our customers. In Canada, we are constrained by strong Privacy laws and, for now, this cannot be done. However, AI and enhanced computing capabilities and efficiencies will bring all these data sets together to do just that: drive insights for companies that will bring them closer to their customers and mitigate their slow death into irrelevancy.
What I have seen in my recent foray with enterprise clients is that there is a desperate need to get on the AI bandwagon and invest millions to figure it out. Companies who want to minimize risk will experiment with ML without really understanding why or how its results will impact current stakeholders, process, job descriptions and employee churn.
For startup companies like ours, who have seen nirvana and know how to carve out a reasonable path to building AI, we are not mired in the complication of redefining all aspects of the business, impending costs and people implications. We just know that TODAY, if anyone doesn't start thinking about AI tomorrow they are clearly at risk of becoming irrelevant. Consider the closing of ToysRUs. There are a multitude of examples of large, once-successful businesses that haven't evolved with the times and we are witnesses to this trickling effect into business obsolescence.
Where I am today...
Start-up is hard. Trying to find that unicorn and staying on for the ride is just as difficult. For AI, this shiny new object, we are at a place where people and companies can still learn and be part of the AI economy. But they should start NOW.
I am now with Salsa AI. We are a non profit organization building a platform for everyone. What is clear today is that there are very few data scientists with the skills needed to scale AI. In 2015, RJ metrics scraped 236 MM and reported between 11,400 – 19,400 data scientists worldwide. These numbers continue to be disputed. Today, demand for data scientists has outpaced demand for engineers or analysts by 50%. IBM predicts that by 2020 the demand for data scientists will soar 28%. Unfortunately, those that can afford the demanding rates for these roles are the Googles, Microsofts, Amazons, big banks and big pharma.
What we also realize is the data science discipline will not be the panacea to solve the world's problems. They are hampered by the narrow scope to solve very specific problems. What they don't see are the nuances of the problem, the potential causes, or how it's identified by the human. They don't see the impacts to the business and how it may be perceived if the company decides to roll it out. They are also inundated, because there are so few of them, with solving big business problems for the sake of profitability. But, who is solving the little issues we face on a day to day basis? How can we make things better and more efficient in our clubs, our homes and in our communities? AI will require people to be the innovators and the problem solvers. Machines cannot yet replace human creativity. We need to capitalize on these awesome ideas.
At the same time, the rise of the machines has instilled an unnecessary fear that people will lose work and purpose. We believe that democratizing access to AI beyond the data science community, into the hands of everyone who wants to create and solve little or big challenges, will address the scalability concern. We also realize not ALL of us are meant to do the math. Access to AI should not be impeded by the ability to understand the machine learning languages. And it doesn't need to be.
Here's the thing: Industry will not wait, and Google wants to supply the AI Demand at scale: We have exactly the same mandate as Jeff Dean of Google, AutoML: “We want to go from thousands of organizations solving machine learning problems to millions.” Caveat, human knowledge and human access are what will effectively scale AI. Machines are not yet able to effectively mimmick the complexities of the human mind... nor should they.
As our business evolves and as we all learn with it, my discussions with enterprise, emerging start-ups and AI enthusiasts in this space will shape how we build our platform. That will also influence the stories that I will publish here. My goal is to profile technologists who are succeeding in AI, and thought leaders who witness gaps in AI's emergence and want to solve them. I can't wait to share these stories with you. Onward!