Not a day goes by in the C-suite where Artificial Intelligence (AI) is not mentioned. AI is top of mind for everyone. With the biggest names in technology telling us all that without AI our lives will plummet into the Dark Ages, it’s clear why AI is top of mind. Many people think of AI as a technology for the back office — “It’s something the IT team will do.” Wrong! AI is a transformational technology that has broad impact across your entire enterprise.
Many of the Communications Service Providers (CSPs) we work with are demanding increased operational intelligence to reduce the cost of their operations and improve customer service. To help achieve these objectives, we work with them to employ highly scalable AI-driven analytics on very complex and distributed networks. The AI-powered analytics are used to predict system failures well before a subscriber is impacted and even prescribe how to fix the issue. This helps them save money by reducing the time required to fix problems, and also reduces the risk of their subscribers being impacted by outages. Introducing AI into the business to save money and reduce costs is not a low-level discussion. This is very often a C-suite conversation that delves into higher level company objectives and ROI. The AI technology is generally secondary or not discussed at all. The business case is the star and AI is supporting the business case.
With examples like the ones above, AI will become part of all businesses in some way. From optimizing deliveries of goods and controlling drones to optimizing surgical robots — AI will be a part of your business and part of your life. But let’s be very clear, AI is not the end-all, be-all technology that we dream of and fear. AI today is very good at solving complex repetitive problems. It learns patterns and identifies outliers and translates that into actions. We experience a form of AI every day when we start our computer in the morning and the cyber security applications protect us from various viruses and threats.
Make a business objective not an AI objective
AI is clearly a transformational technology. Implementing AI in your business is surely something to consider. Considering how to use AI can be challenging, however. Just calling your senior leaders into a room and telling them to implement AI won’t work. Look at your corporate objectives. They will include objectives like reduce costs, improve efficiency, grow customer base, etc. As you develop strategies to address these objectives, there are areas where AI and analytics can help.
Creating a plan with your team and vendors is the only way AI will be effectively introduced into your business, as your team defines the goal for your program. A couple of examples:
– “We need to reduce the cost of outages by $10 million in 2020.” It’s critical to define your success milestones for the year.
– “We will select vendor partners in 1Q and see a 15% reduction in outages in 2Q, followed by reducing costs by $6 million in 3Q and achieving our objective of reducing costs by $10 million by the end of 4Q.”
These, of course, are common sense but it’s critical to define success at the beginning of the project. It will challenge both your staff and vendors and without these milestones, transformations will never be realized. There are so many examples of this, though the word “transformation” is unspeakable in some boardrooms. If you look at the postmortem of failed transformations, there was often a large objective but very impractical interim success steps that created a set of scenarios that forced failure. They created constant points of contention among the staff and vendors, and the leadership teams were constantly nervous as the cost of the programs rose and there were no clear evidence of success along the journey. Mutually agreed-to success milestones are key for any program, and especially key for AI-oriented initiatives.
AI ain’t free
There is tremendous positive ROI and benefits that can be gained from AI, but it isn’t free. The cost of collecting new data sources, computing, and storage can be expensive resources. When embarking on your AI and analytics program, here are a few items to consider:
- Building our own. Always consider build versus buy and the long-term value of both options.
- Specialized applications versus general-purpose AI. General-purpose may be cheaper on Day One, but costs could escalate over time and the levels of effectiveness may be less than optimal.
- Cloud versus on premises, or hybrid. Consider your costs over many years as these options could impact your ROI in unexpected ways.
Like any IT project, AI-based solutions have costs and benefits and careful planning will set you up for success and ensure you achieve your business objectives. Many of our CSP customers see 10s of millions in ROI from AI-based analytics. However, AI solutions aren’t magic, and they aren’t free. There is cost in the form of capital, opex and human resources. Your company must consider how AI will benefit your business but go in with your eyes open. As with many leading-edge technologies, the pros outweigh the cons, and with focused business objectives, planning and solutions you will be happy with the results.
Author Bio
Ben Parker is CTO of Guavus, a Thales company and pioneer in AI-based analytics for CSPs. He is responsible for Guavus’ strategy and expansion into new markets, including IoT industry verticals. Prior to Guavus, Ben was the Director of Network Planning with Verizon, where he led a team responsible for long-term network planning for the Verizon Wireless network. In 1997, Benjamin joined Sprint and began his career as a Network Systems Specialist. He quickly earned promotions to more visible positions, which include Test Engineer, Principal Network Design Engineer, Distinguished Member of Technical Staff, and Sr. Manager of Technical Staff. Following his tenure at Sprint, he held the position of Vice President of Engineering at Oakley Networks.
In addition to his extensive professional experience, Ben holds a BS in Environmental Science from the University of Nevada at Reno and a Master of Business Administration from Benedictine College. His post-graduate training includes Electrical Engineering and Telecommunications from UC Berkeley, and Project Management from George Washington University. Additionally, he has authored or contributed to more than 75 patents all in the area of telecommunications.
