Artificial Intelligence is rapidly moving to the center of enterprise thinking about Digital Transformation. Gigster has delivered over 1,500 digital transformation projects, with over half of our projects now involving AI algorithms. At the Google Next conference in San Francisco, Gigster announced a partnership with Google Cloud AI and held a roundtable discussion on how to drive business value from AI. This blog post grew out of that discussion. *“We are now in the period where it is critical to put Machine Learning inside all the apps, just like in the 1980s we were putting SQL inside all of the apps.” — Frank Chen, Partner at Andreessen Horowitz*
Frank explains that machine learning has transitioned from a feature that was nice-to-have to one that is now a must-have for enterprise applications. History has a way of repeating itself. In the 1980’s, SQL enabled analytical insights that improved the value of decision-making across the board for the enterprise.
Today, machine learning is enabling predictive insights that accelerate the pace of decision-making across the enterprise. Before SQL a large burden was placed on the developer to “navigate” data in order to get the desired information. Relational databases and SQL relieved the developer of this task, providing an accessible path to insights.
Today, modern AI techniques like deep learning are sparing the developer from hand crafting features that describe the data. Moreover, sophisticated AI frameworks and APIs have made capabilities that power AI at Google, Facebook and Amazon available to everyone.
“Most of the value created with relational databases was captured not by Oracle but by the Fortune 500 companies who used Oracle. AI will be the same – yes big tech will benefit, but most of the value will come as enterprises transform their business with AI” — Frank Chen
Much of the focus on companies benefiting from AI is on Google, Microsoft and Amazon, as well as companies like Databricks, which inform the market about what is possible and provide an enabling AI infrastructure.
This is akin to only focusing on Oracle as the company that would benefit from relational databases. However, as with databases, the bulk of business value creation will occur when enterprises take the potential of AI to heart and use it to deliver new business models, like self-service claims adjusting in insurance, real time Know Your Customer in finance and self-driving customer service in retail.
Like the SQL revolution, a significant friction point for adopting machine learning is the talent shortage. Gigster customers and partners report that hiring a machine learning expert in a big tech center like San Francisco or Boston can take three to six months – hiring these skills in other locations can be next to impossible.
The other challenge to overcome is organizational maturity around best practices and norms. The story of databases teaches us that there is hope. Over time as more work is done by the underlying infrastructure, managed services and AI frameworks, the technology will get more accessible widening the pool of available talent.
In addition, ideal structures and organizations will emerge as more companies deploy AI solutions and trade best practices.