Digital Technology
Sep 08, 2017

AI for the Enterprise:  Where to get started and how to pragmatically go about it

With so much written on AI today, most enterprise leaders are aware of its transformational benefits, but many still wonder where to get started.  And how to do it pragmatically.  This post describes the type of AI which is ready—and optimized—today for enterprise deployments and provides guidance on where/how to best get started.

First, Popular Science and Wikipedia have provided definitions for the different types of AI.  “Narrow” AI performs one narrow task, as opposed to Artificial General Intelligence (AGI) which seeks to perform any intellectual task as well as a human.  AGI leads to the notion of Artificial Super Intelligence which can be thought of as an intellect that is much smarter than the best human brains in practically every field.

While AGI garners more headlines and raises concerns about future control and governance, enterprises should focus their investments on narrow AI.  Narrow AI is ready to deploy today.  It is this type of AI which some enterprise leaders are using today to increase their competitiveness and market value.

Where to get started

My earlier blog on AI for the Enterprise described the significant differences between AI developed by the tech titans (e.g., Apple, Google, Facebook, Amazon, and Microsoft) for B2C use cases and the unique enterprise requirements for B2B use cases.  While keeping these enterprise-specific requirements in mind, leaders should focus their initial AI projects around the following principles:

  • Stay close to the core

Narrow AI is finding applications across the value chain, but some parts of the value chain deserve more attention than others.  Customer service, operations and product development are good areas to start since they are typically core value chain elements across many verticals.

For financial services firms, core means focusing on operations and customer service. For consumer packaged goods, core means operations and sales/marketing. McKinsey Global Institute research shows leading sectors are not only deploying AI in core parts of their value chain, they are also deploying it in more parts of their value chain.

For example, leading financial services firms are using AI today to transform one of their core operations—financial spreading.  To see the impact of AI, consider:  a human with four years of experience can spread financials in over 60 minutes with 80% accuracy rate and a cycle time of over one day (with no audit trail).  AI-based systems spread financials in less than 15 minutes with 99+% accuracy and a cycle time of less than four hours—all with an audit trail.

  • Grow revenue; don’t just cut costs

Enterprise leaders in AI adoption tend to use AI more frequently to drive revenue and market share growth vs. experimenters who are behind tend to focus their AI on cost reduction projects.  Research also shows that the more enterprise leaders use and become familiar with AI systems, the more they invest in projects to drive growth.

For example, one large manufacturing company has used AI to transform their quote generation process.  They are leveraging computer vision and other AI-based technologies to better enable their distributors to read architectural blue prints, which has reduced quote turnaround times by almost 90%.  As a result, this narrow AI solution increased their quote win rate and helped drive revenue growth and market share gains.

  • Transform end customer experience

Enterprise leaders recognize narrow AI as an opportunity to not only streamline core operations and drive revenue growth, but—perhaps most importantly—to create new winning end customer experiences.  Front-end client-facing operations are ripe opportunities.

For example, in the Life Sciences vertical, pharmacovigilance is focused on managing adverse effects reporting for new drugs—in order to improve drug safety. AI technologies are being used to extract unstructured data, read and make sense of it, perform complex analytics and continually learn. Narrow AI has transformed this formerly labor-intensive process to not only save hundreds of millions of dollars, but to also improve the consumer experience and—most importantly—to potentially save lives.

How to go about it

Enterprise leaders who focus their initial AI projects in the above areas will do better than others. In addition, they will learn that successful AI adoption requires buy-in by the executive suite to generate the momentum needed to overcome typical organizational inertia.

Lastly, one other key element for successful AI projects is domain knowledge. The challenge of applying AI to the Enterprise is not one of technology; rather, it is one which requires contextual understanding of and deep business understanding.  This domain expertise is needed either from a Center of Excellence (COE) or from a trusted partner. One AI-based solution provider put it this way“The key to our success is domain knowledge: I have a team of experts.”
While AI technologies grow ever more sophisticated, one cannot overstate the importance of experience and the human element.  Since digital transformation is more about mindset than technology, your company needs to ensure this domain expertise is available to optimize the success of your AI projects.  Doing anything less is foolish.

About the author

Dan Glessner

Dan Glessner

Vice President, Digital

Follow Dan Glessner on LinkedIn