Big data and Artificial Intelligence (AI) complement each other. Businesses grapple with the proliferation of data caused by digitalisation. AI offers improved opportunities to make use of such data, to derive real-time actionable insights from it. But the success of AI rests on huge volumes of high-quality data. AI algorithms learn from existing data and apply what they have learned to new data. Here are some use cases that make explicit the multiplier effect of Big data – Artificial Intelligence conjunction.

1. Gleaning Structured Data

Over 70% of enterprise data come unstructured. Finding the “usable” or relevant data fit for Big Data analytics is tricky. The sheer volume makes it impossible to structure unstructured data manually. Machine learning culls the wheat from the chaff.

Exasol’s AI algorithm extracts structured data from non-standard formats, making it ready for analytics. In a real-world business, each invoice looks different. Different names and languages make normal computer programmed extraction impossible. The team led by the CIO trained a model that learns from scanned invoices and historical data, to extract the required fields. The algorithm created a model that structured the data by just scanning new invoices.

Another big application of AI in Big Data is to make sense of emails. The average knowledge worker sends and receives 120+ email messages a day. VKB, a Germany-based insurance company developed an input management tool to identify the topics and sentiment. The tool unearths such information from unstructured text in incoming emails and routes emails.

2. Streamlining Complexity 

Big data is often a result of complexity and bureaucracy. Applying AI to Big data allows enterprises to cut red tape and improve outcomes.

The best manifestation is back-office operations at banks. Banks struggled to process the large and complex data sets involved in key processes such as loan processing and “know your customer” process. Implementing robotic process automation powered by machine learning algorithms delivers significant savings on time and costs of such processes.

Another area where Artificial Intelligence meets and resolves complex Big data is in mortgage processing. Unlike human agents, AI algorithms and the associated processing infrastructure does not get tired or cause mistakes. AI automates the critical path process, reducing the processing time for the mortgage from three to four weeks to about two days. This increase in throughput makes the home-buying experience less stressful and also faster for the buyer.

Using AI at the point of sale provides borrower self-service. AI enables real-time processing. Borrowers perform the bulk of the loan application process, and later loan maintenance without getting in touch with an agent. The AI-powered bot automates Big data processing, to enable such self-service.

At a macro level, DHL’s Global Trade Barometer uses AI to evaluate 240 million variables from operational logistics data and advanced statistical modelling, to generate a monthly outlook on the state of the global economy.

3. Assisting the Customer

AI meets Big data to unlock useful information, and meet customer needs. Bots use AI to scour through Big data and meet customer needs quickly and accurately.

OSHbot, the customer service bot developed by Lowe, the American home improvement retailer, assists customers. The bot uses deep learning algorithms, natural language processing, and computer vision to process and identify the item requested by the customer. The bot then accompanies the customer to the location of the item.

KLM Royal Dutch Airlines BlueBot delves through the maze of Big data, to help customers book flights. It goes on to offer a host of value-added advice including what to pack for a trip.

CircleBack’s Artificial Intelligence engine processes billions of data points to ensure accuracy and timeliness of contact information.

4. Making Better Use of Video and Voice Assets

Until recently, enterprises were clueless about how to include audio and video files to Big data analytics.

Neuro-Linguistic Programming (NLP) automates voice-to-text transcription. This enables the application of speech analytics to audio files, to derive actionable insights from voice data. Enterprises playing the message “this call may be recorded for quality assurance and training purposes”, applies such Big data analytics.

Companies store call recording data for manual review and compliance reasons, sometimes for seven years or even longer. AI resolves the storage challenges associated with call recordings and other voice data. AI transcribes voice recordings in real-time and spares the huge expenses to store uncompressed audio.

In video content, AI delivers advanced metadata enrichment, to optimize video libraries. Speech-to-text technologies improve the accessibility of video applications and enable live-feeds. For instance, AI generates live and on-demand automated captioning for video streams.

Uber uses satellite imagery from DigitalGlobe to develop advanced mapping tools, to improve pick up precision and optimize ride routes. DigitalGlobe’s satellites decipher new road-surface markings, lane information in real-time, and updates live traffic info to its vector map.

5. Accelerated Processing

Application of AI speeds up deriving benefits from Big data.

Conventional keyword search often returns low-quality results. Knowledge workers today spend anywhere between 15 minutes and an hour per day searching for business information in outdated database systems and corporate intranets past their sell-by-date. AI applies techniques such as ontologies, automatic clustering, and visual recognition to reduce search time by 50% or more.

Helio-graf, an AI-based solution developed by Washington Post, generates short reports from the data and narratives provided by human reporters. The tool generated around 300 short reports covering the Rio Olympics, freeing up journalists to do high-value work. A spin-off benefit is accuracy in financial reporting, free from human errors.

Leverton’s AI-based platform simplifies processing and management of real estate contracts. Natural language processing classifies and reviews contracts written in complex legal language. The tool processes contracts extending to hundreds of pages in a fraction of the time taken by human experts to do the same task.

Not too long ago, enterprises viewed Big data as a challenge rather than as an opportunity. Artificial Intelligence enables better insights and converts Big data into an opportunity.