Five Thoughts on Analytics, AI and Automation

Companies that leverage analytics and AI to make data driven decisions have a significant competitive advantage over their peers. [1]

The flywheel effects of digital platforms are creating dominant players that are not only transforming their core business, but are also launching new businesses in adjacent categories. At the center of the platform is the generation and ownership of proprietary data. As the volume of data continues to exponentially grow, companies that are able to obtain, process and interpret the data are able to understand their customers and are capturing the lion’s share of growth. Case in point Amazon, as the e-commerce company was able to understand who their customers were, what they wanted, where they lived and how they behaved, they were able to parlay these insights into new products and services. Amazon is now the number three player in OTT streaming reaching 33 percent of households that use streaming services in the US. [4]

Market share of OTT streaming services [4]

The concept of AI is nothing new however there have been several technological breakthroughs that have been driving deployment. Three factors causing this are the understanding of machine-learning algorithms, increasing computing power required to process large data sets and the collection of massive amounts of data. In specific cases today, companies are able to substitute labor with more efficient and safe alternatives. Rio Tinto was able to deploy self-driving haul trucks at their mines in Australia to achieve 10-20% increases in utilization. Google applied DeepMind (their machine learning AI) to their data centers to cut energy consumption by 40%.

The ubiquity of AI and automation is fueling public concerns about the effect on employment rates and availability of future work. [2]

Concerns surrounding AI and automation are the potential to affect tectonic shifts in the labor market resulting in large scale unemployment. Let’s be clear, this position is nothing new. Similar concerns were echoed during the industrial revolution where workers moved from farms to factories and service jobs. Though it was not obvious at the time, this change resulted in the creation of new types of work and jobs that were unforeseen at the time. Things may not play out the same this time around however an interesting thought is that these technologies might actually help the labor market function better. Similarly to how we were able to develop better, targeted and personalized ads through platforms like Facebook, the use of professional social networks like LinkedIn to improve the matching of workers with jobs that they enjoy and are best suited for may contribute to an effective, efficient and content work force. Data-driven hiring and placement of workers can raise labor participation and working hours while reducing turnover and unemployment.

Automatability across economies [2]

We’ve had guest speakers from large corporations that are being disrupted (Jeff Wesler, IBM Research) and start-ups that are disrupting (Simon Chan, Prediction IO). What is clear is enterprises that hope to be successful should adopt the following mindsets:

• Test new ideas, implement leading technologies and quickly scale the ones that show early signs of success and adoption
• Reimagine business models and business processes to make use of data driven technologies
• Continue to invest in human capital through training and integrating their workforce with machines and computers

Companies need to be aware of unintended consequences of data driven decision making on society and adapt to changing social and political climates. [3]

On April 21^st, 2016 Bloomberg published a piece titled “Amazon Doesn’t Consider the Race of Its Customers. Should It?” The issue with the data driven decision making process is it eliminates the human element that’s important for society. Companies need to underpin their business decision making with social logic to provide the best and most robust business model. In the case of Amazon, Bloomberg overlaid the areas where Amazon offers same day delivery with racial demographics. What they found was that many African American, Hispanic and minority communities were excluded from same day delivery. Based on data from consumer spending patterns, profitability and other performance metrics Amazon only offered the service in particular areas. Though it was not Amazon’s intention to underserve a demographic of the population, it was still an unintended consequence that caused societal and political controversy. Shortly after Bloomberg published the article, Amazon extended coverage of same day service to many urban zip codes that were predominantly African American, Hispanic and minority neighborhoods.

Zip codes where Amazon offers same day service [3]

Map of Atlanta Georgia overlaying population, racial demographics and Amazon’s same day service areas [3]

• https://go.forrester.com/wp-content/uploads/Forrester_Predictions_2017_-Artificial_Intelligence_Will_Drive_The_Insights_Revolution.pdf
• http://www.mckinsey.com/global-themes/digital-disruption/whats-now-and-next-in-analytics-ai-and-automation
• https://www.bloomberg.com/graphics/2016-amazon-same-day/all-cities.html
• https://techcrunch.com/2017/04/10/netflix-reaches-75-of-u-s-streaming-service-viewers-but-youtube-is-catching-up/