Turning Massive Data into Meaningful Insights

January 25, 2019 Faiz Ahmed

 

Applying AI and Machine Learning to Big Data for Better Client Outcomes

In 1992, there were just one million connected or “data-generating” devices in operation globally. By 2012, just 20 years later, there were 8.7 billion connected devices. If you think that pace of growth is fast, consider that by next year, there will be 50.1 billion connected devices sending and receiving data every minute of every day. That’s a heck of a lot of data, and it’s not just coming from our mobile devices. In fact, mobile phones will comprise only about 10% of all connected devices by 2020 — with other devices such as sensors used to gather shopper information, tablets recording purchase transactions and GPS location devices producing massive amounts of data across every business and industry.

 

 

 

In developing countries, the average adult now leaves behind 11 hours’ worth of a data trail each day.  Globally, and with all of our connected devices, more data has been produced in the last two years than in the entire history of the human race before this time. This is what we call “big data” and it holds great promise in the world of legal, risk and compliance.

 

 

So, how do we derive value from all this data?
At last year’s Continuum conference, I outlined three primary steps to derive value from big data sets. Let’s take a look at each step in the process, and along the way we’ll use an example of an actual business application we’re working on right now with one of our telecom clients – cell tower leasing.

Step 1: Frame the problem or opportunity.
If you’re looking to gain value from your data, it makes sense that you first define what you’d like to know. What information are you responsible for knowing and identifying — and what will you do once you have the knowledge you seek? Framing the opportunity and problem is important because the more specific and clear the description of the need, the more likely the right data sets will be identified, collected and processed. 

We recently had the opportunity to work with a telecom company that is working to provide 5G service to its customers. Telecom companies are in a race to provide 5G service. Doing so requires that they hang updated equipment on all of their cell tower sites and also build new towers. In fact, 5G service requires as much as 16 times the amount of equipment than is hanging on towers today. But telecom companies don’t typically own the cell towers or even the land on which the towers sit. Rather, they lease both, and the complexities associated with managing the leases are enormous. A telecom company may have as many as 50,000+ site leasing contracts with tower and land owners.

So framing out the problem starts with some important questions. For example, on which of the thousands of towers does new equipment need to be installed?  If new equipment is required or a new tower needs to be constructed, does the applicable lease contract allow for it — or does it have to be renegotiated?

Step 2: Identify and Process the information
Visualizing information is key. Business people need to be able to see the processed information cleanly. They need to be able to examine the data at a high level, understand it, look at it in different ways, and identify patterns. From this, initial insights and theory are derived. 

But first, identifying and processing the sources of relevant information is necessary.

In the cell tower leasing example this critical step includes gathering all of the thousands of lease contracts. It also requires gathering an inventory of the updated equipment that is required for each cell tower location, and the costs associated with each change or addition of equipment. 

Once the company has identified and gathered the relevant data sets to address the business issue, and can cleanly present it to business and legal units, a data science team begins its work. This involves identifying the algorithms required to apply a machine learning-based solution that will secure specific information and insights needed to make business decisions.

Step 3: Apply AI and Machine Learning

It’s important to know that machine learning is a form of AI that applies to big data problems.  It facilitates an automated set of processes that learns from the data and adapts. The application of AI identifies patterns and even makes decisions with minimal human intervention. The enablers of machine learning are data scientists who work with business people to understand goals, objectives, and the knowledge sought, and then use algorithms to build models that uncover connections in the data. Then, as human users (including customers) give feedback, the system becomes more and more accurate. In the case of the telecom provider, the application of machine learning AI is the only way to truly pinpoint which of the thousands of cell tower locations can be updated immediately with the required equipment, and which ones require the lease contracts to be renegotiated before the updates can occur. 

 

 

In the context of the cell tower leasing example, we can apply a set of provisions to the data such that our telecom provider client is able to quickly and easily identify all the leases that are impacted by the upgrade to 5G, based on key rules that human review alone would surely miss. It’s not about removing humans from the equation.  In fact, the human feedback on initial results help the system learn and find more exact results. Moreover, this helps humans identify the right information much more quickly.  In what is a great global race to provide 5G service, the time savings are of critical importance to our client’s business and their end users’ satisfaction.

Summary
Using machine learning provides immediate, reliable results and represents a massive savings opportunity in terms of time and resources. With practical applications across every area of business including contract compliance, data privacy, lease administration, content moderation and many more, we are helping our clients uncover more value from their big data in less time than ever before. Together, we’re creating more individualized, immediate and intelligent interactions at every step of the customer journey.

 

Learn more about Conduent’s Legal and Compliance Solutions.

About the Author

Faiz Ahmed is General Manager of Conduent Legal and Compliance Solutions

More Content by Faiz Ahmed
Previous Article
Beyond eDiscovery: Applying AI over a Platform Model to Streamline Legal Operations
Beyond eDiscovery: Applying AI over a Platform Model to Streamline Legal Operations

Next Article
GDPR is here and CCPA is coming — is your legal and compliance operation prepared for Subject Access Requests?
GDPR is here and CCPA is coming — is your legal and compliance operation prepared for Subject Access Requests?

When the General Data Protection Regulation (GDPR) went into effect in May across 28 countries in Europe, n...