Choose your language:

Australia

Germany

Hong Kong

India

Ireland

Malaysia

Netherlands

New Zealand

Singapore

Sweden

United Kingdom

United States

Next-gen business process transformation through AI

3 real use cases

April 10, 2018 | By Ram Palaniappan

artificial intelligence automates some tedious tasks for workers in an office

Artificial intelligence and machine learning have long shown extraordinary promise for reducing business inefficiencies and speeding up innovation—but their broad accessibility is relatively new, leaving most IT and business leaders without much of a roadmap. 

Our Data Insights team is working hand-in-hand with several Fortune 100 organizations on AI and machine learning applications they can use now, and I’d like to share some of the most successful practical use cases that apply to broad business and IT processes in need of optimization. 

IT network support and monitoring

The massive quantity of data most businesses are collecting from network operations, security operations and privacy threats makes it challenging for security analysis teams to manage its volume and mine it for relevant threats and potential outages. This data processing, synthesis and pattern analysis may be the single most useful AI application for most network operations units today.

We’ve built an AI platform that processes and synthesizes data to identify patterns that lead to root cause analysis. This is augmented by a self-service autonomous chatbot that helps IT support for issues like security flaws, network outages or applications access issues get accomplished more autonomously. Tech support teams handling user issues one at a time might take a long time to recognize these patterns, but the AI bot synthesizes data in real-time and performs analysis to identify the root causes and resolve them. 


Back office operational efficiencies

We’ve worked with customers to augment business processes like procure to pay, order to cash, hire to retire and record to report with AI algorithms and closed loop automation. For example, we worked with a multinational operation to reduce expense approval processing time. Instead of submitted reports going through a 100-person shared services team for manual first inspection, a bot examines and automatically approves 40 percent of the expense sheets. It routes another 30 percent back to the expense submitter for additional details whenever needed. Only 30 percent of expense requests now require manual scrutiny. 

The billing analysts also spend a huge portion of their time in categorizing invoices and assigning them to the right billing category, and checking the invoices against policies and authenticity of transactions for issues—tedious and time-consuming work. The AI bot has taken over those functions, reducing the analyst’s workload by about 35 percent—the equivalent of 35 full-time workers. The organization has used this extra capacity for other areas, and retrained staff to meet a higher volume of billing.  

It’s easy to see how other types of organizations could use this approach to increase the speed and capacity of back office functions. 

Manufacturing quality 

Manufacturers of complex products rely on multiple components from multiple suppliers, and one defective part will cause the whole product to fail—which is expensive and sometimes dangerous. The costliness of these problems, especially for high-end electronics like smartphones, means companies place strict quality controls on their suppliers. Those suppliers, in turn, conduct rigorous testing, which can be incredibly time-consuming. 

For example, we worked with a company that built high-end sensors that had to undergo 100 hours of testing for each sensor. This created a large bottleneck that slowed down production. We built an AI system that visually inspects the chip and correlates it to test results and quality parameters captured through deep learning algorithms. We trained the AI using six months of rich data assets, which allowed our solution to predict the parts that would fail early in the testing process. This allows the manufacturer to halt testing on parts likely to fail, dramatically reducing the bottleneck and increasing speed of overall production. 


A model for piloting practical AI and machine learning projects

TEKsystems brings the power of AI and machine learning into regular business applications and processes through augmented intelligence to assist organizations with tasks that are time-consuming or require processing of large quantities of data. 

This is made simple through the smART Machine Framework we developed in-house. It incorporates natural language processing and machine learning capabilities through an open source platform. The framework allows companies to accelerate proof-of-concept projects and scale quickly to full-fledged AI applications. On average, the return on investment for AI-powered solutions is anywhere between 6-10X. 

Our smART Machine Framework helps clients run quick, controlled proofs of concept, prove business value to top management and quantify ROI before making a full investment. 

Because the ability to use AI tools for business problems is so new, I advise clients to work with consulting services companies that have bleeding edge technical capabilities combined with domain expertise. This can help you identify the right use cases and right platform components. Then take a lightweight approach to testing them, using a proof-of-concept model.

Do you have a business problem that you think an augmented intelligence might solve? TEKsystems can help you assess your processes and use our smART Machine Framework to pilot, develop and operationalize AI applications—before the competition catches up. Contact us to start exploring how AI can help you today.