Our client, a multinational telecommunications services company had long since unearthed the single biggest cause of service disruptions – network fault. They now sought a way to be able to predict and prevent network fault in a bid to avert service disruptions.
Infosys sought to help address this challenge.
Infosys started with the task of knowledge curation beginning with analysing the ADSL service layer data. Typically this data has multiple connection property information like attenuation/loss, code violations, upload/download rates and re-initializations that can serve as early indicators of network fault.
Leveraging the Infosys Information Platform, we then analysed the service layer data of customers from periods where they did not report any fault to arrive at a “control signature” using signal processing and statistical techniques on connection properties. (Such as attenuation/loss, code violations, upload/download rates, and re-initializations) The “fault signature” was then computed from the connection properties in the days leading up to the reporting of a fault. Armed with both signatures, a statistical model was used to determine the “signature” of a connection with propensity to exhibit a probable fault. In this manner, the impending fault report for the forthcoming week was published. This model can be applied to a single connection, a DSLAM or a small geography with equally successful results.
Of all the network lines this telecom has 3% of lines were identified to have a fault probability greater than 60%, in a span of 3 weeks. This gave them a head start to segment and then prioritize these lines to fix issues before they occur. This model is being fine-tuned and used to find a segment of the population with the highest likelihood of fault reporting and to actively intervene to mitigate the weaknesses of the network.
A firm engaged in the manufacture, sale, installation and service of financial self-service transaction devices (like ATMs) was looking to curtail the burgeoning costs of ATM maintenance and improve customer service.
Infosys partnered with the business to help tackle this challenge.
Knowledge curation and analysis was the first step undertaken. 4 million records related to ATM faults from 8500 ATMs were studied.
The tickets raised against faults were plotted against the ATMs, to determine the ATMs with maximum outages. Of these, the ATMs that consumed most maintenance time were identified. The same plotting was also established for defect types to determine the type of defects that take the most time to resolve. Interestingly, the largest number of defects were of the type that demanded the least time investment to fix. This exploratory analysis was continued to also determine the geography with the most breakdown incidents.
We then leveraged Infosys Information Platform to derive foresight from these insights using an analytical model to predict the failure of ATMs taking into account the key factors that lead to such failure. Some key drivers of faults were identified as:
The logistic regression algorithm of the SPARK machine learning library was used to make the predictions.
Now it was possible to easily determine – say, those ATMs with 90% propensity to fail.
This predictive model, when implemented, is slated to reduce the downtime of ATMs by 10%, creating an increase of 15% in transactions that might otherwise have been lost because of faulty ATMs.
One of the world’s largest mining companies has a number of autonomous and unmanned trucks operating across various geographies. Breakdowns of these vehicles impacted the entire supply chain with significant consequence on the business. This company also plans to scale the number of trucks over a period of time. The trucks have approximately 200 sensors fitted to stream information back to the system. It is essential that the information coming from the sensors is used in real time to derive insights around the functioning of the trucks and prevent it from failing.
Infosys channeled the entire stream of data from the truck sensors into the Infosys Information Platform data lake through Spark streaming. The streaming rate was very high with a throughput of 27,000 messages per second. Key levers that impact functioning of the trucks and cause failures were identified in consultation with subject matter experts. These included noise level, thermal variations, temperature and tyre pressure to name a few. A time series analysis was made on the entire volume of data, emerging from the sensors, and if these key levers shot above threshold deviation (as prescribed by subject matter experts), the truck was predicted to be predisposed for failure. The results were clearly displayed on an easy-to-read dashboard. The snapshot on the dashboard was foresight on display emerging from the insights contained in 300 million live records which were analyzed in less than 4 seconds.
A large freight railroad network in North America was keen to tackle the problem of excessive unwarranted braking events of its locomotives. These train stoppages (braking events) were leading to reduced velocity and locomotive utilization. This was negatively impacting the organization’s operating ratio and customer service arising from their inability to adhere to published operating schedules.
Infosys took the initiative to help find the correct insights that could lead to a solution to this challenge.
The Infosys team took a closer look at the knowledge that already existed within the company with regard to locomotive brake data. This was evaluated in combination with engineering characteristics, wayside data streams, maintenance zone information, environmental and atmospheric data and PTC signal data. This analysis provided a 360 degree view into the landscape contributing to the braking events.
Leveraging Infosys Information Platform, and a delay event prediction model using R - a free open source statistical language, along with AWS Cloud for reporting, we were able to establish, through a Pareto analysis, that the signals, speed restrictions and switch alignment were the top contributors to the braking events.
We also established that should there be an active data feed live streamed into the model. This methodology will ensure that alerts are generated when delay thresholds are surpassed and in turn trigger delay alerts for downstream stations.
Infosys Information Platform is fully equipped to manage real time PTC signal data aggregation, which can run into several 100 terabytes, and execute real-time predictions as well.
The client, a global provider of products, services and solutions to industrial and commercial users of electronic components, engaged Infosys to steer a large scale global business transformation program spanning across 75+ business processes and 45+ modules including supply chain management, CRM, distribution and finance. To avoid disruption to business and address all showstopper issues before the 3000 odd users start business activities on the new platform, the client sought to conduct a health check scoped to include checks for:
Using Infosys Automation Platform, the Infosys team successfully automated the health check process to a No Touch Program. This meant automating activities like logging into remote client machines, opening Oracle instances, checking various forms and generating reports.
Infosys provides application management services, in a managed services model, for a large telco in Europe. The scope of work includes supporting business critical applications such as the order management system that processes over 5 million orders every year, along with a network provisioning system used for order fulfilment. In an increasingly competitive environment, where customers can change telecom providers with a click of a button, fast and efficient customer service is critical and our client sought to improve it.
Infosys leveraged Infosys Automation Platform (IAP) to automate various IT service management activities such as service request fulfilment, incident management and change management thereby making customer service and IT operations faster and efficient.
Infosys used IAP to create intelligent software robots capable of resolving customer service requests such as issues around order fulfilment with blocked orders and in scheduling field service technician appointments. This is helping the client service customer requests faster and efficiently.
Better customer experience with faster resolution
Our client is a global leader in selling, marketing and distributing food products to commercial establishments. With a diverse product line providing solutions for meals consumed away from home, they operate over 196 distribution facilities serving over 425K customers. Infosys provides them IT infrastructure support services, application maintenance and support services. Their IT landscape includes diverse technology stacks such as Oracle databases, SQL Server databases, Middleware, SAP BI4, SAP R/3, WebLogic and DWBI, and they sought a solution to ensure their IT SLAs are met even during peak workloads.
Infosys team analyzed historical support tickets using Infosys Automation Platform (IAP). Analysis helped identify patterns in tickets and distinct opportunities for automation. Infosys leveraged Infosys Automation Platform to create intelligent software robots who resolve incidents. These robots are now helping address peak workloads and meet SLAs.
A large European turbo-machinery manufacturer sought to automate the machining processes for its turbine blades– an otherwise time-consuming process when executed manually.
Infosys recognized this is one of the most complex knowledge-based engineering engagements requiring in depth knowledge of computational geometry, software techniques as well as a firm grasp of domain knowledge in turbo machinery. The team successfully executed the development of one of the most business-critical applications, using customized CAM software, for designing, rendering and generating manufacturing data for blades in the manufacturer’s turbine.
A global leader designing and manufacturing diesel engines and diesel power generators sought to improve the efficiency of their generator design process which involved multiple component integration and complex engineering.
Infosys helped to automation the design of the generator set chassis subsystem by seamless integration of analysis and detail design using a commercial third party CAD /CAE system.
A large oil & gas pipeline operator based in Canada wanted to better plan the routing of its oil commodities through client networks.
Infosys took on the challenge to design a knowledge-based engineered system to plan the assignment and routing of shipper nominated oil commodities on client networks at optimal cost, uncompromised quality of delivery, and streamlined operational efficiency while being fully cognizant of constraints related to capacity and maintenance of pipelines. Infosys adopted a hybrid approach leveraging the knowledge of an expert allocator, robust mathematical models and operations research techniques to achieve operational efficiency of the volumetric engines.
The service agents at the contact centre of a major US-based insurance firm needed individual sign-ons for the various applications they logged into every day. These applications, cluttering their desktops, made it cumbersome for agents to find the right information quickly. Notes captured manually led to errors and reduced scope for reuse. Often, time-consuming and repetitive searches were performed across multiple systems. Our client sought a solution to this time and effort drain.
The team proposed to leverage AssistEdge. The platform’s Sign-In Manager provided a single interface to log into multiple applications. The unified desktop feature non-intrusively integrated heterogeneous applications on a single screen. The AssistEdge electronic call log helped configure call notes, so service agents could easily refer to these at a later point. Workflow and process automations enabled single click search and update.
Our client, a large global investment bank based in Australia offering multiple financial services and with operations in over 30 countries, had poor spend visibility into procurement expenses. The bank lacked an integrated application environment and hence had little price benchmarking, reconciliation and sourcing capabilities. Their supplier base was fragmented resulting in inefficiencies in managing spend and operations. The bank sought to bring in a better managed procurement system.
Leveraging ProcureEdge, the bank redesigned their purchasing framework to route procurement through the appropriate purchasing method thereby reducing maverick spend and enhancing compliance. Process and technology transformation of sourcing and procurement within the bank, introduction of helpdesk tools and payment operations tools improved the other related processes as well.
A large sports goods and apparel manufacturer, partnering with a number of retailers and franchisees, needed to closely monitor and manage their view into secondary sales and inventory for the entire supply chain. The retailers and franchisees had varying levels of technology maturity and hence issues in data transfer and collation across the chain were not uncommon. The firm was keen to get deeper visibility into their operations and supply chain with a robust data exchange platform.
Leveraging TradeEdge, the company automated extraction of data both at a pre-defined and configurable time. Business validations were instituted to authenticate the data for frequent issues and alert the retailer before the data flowed into downstream systems. Systems were also put in place to ensure the data was then further validated and signed off by all stakeholders on a monthly basis.
Emerging markets - growth engines of the new global economy – are characterized by multiple tiers of distribution, unique in the challenges they present as also opportunities. Our client, a consumer goods leader seeking to acquire their next billion consumers, was looking to consolidate their position in these markets. The challenge, of course, was that while collectively these markets are bigger than many modern (or big box) retailers in the West, their scale breaks down at an individual market level. Getting to know each channel partner (or distributor), what sales looked like for each of them and understanding the service levels maintained were the first foundational steps to be taken to realize their vision.
When Infosys took on the challenge to help the client realize their aspirations, we were quick to recognize distributors’ deep seated reluctance to share information. We also took into consideration their low process and IT maturity, before proposing a two-pronged approach. First, to create business incentives for distributors who shared data. And second, to refrain from mandating strictly defined data sharing formats. To make this dual strategy work, Infosys built a data exchange platform that could easily cleanse, configure and map data, in any format, into suitable output formats. The platform was hosted on the cloud, with the ability to scale globally and required no installation to be made at individual distributor premises. Further, Infosys leveraged its global presence to set up help desks to follow-up with distributors locally in the event of data not being shared at a pre-agreed time (which also triggered alerts from the platform). This helped build a robust global data acquisition service where Infosys took the bottom line to drive compliance and make quality data available on time.
Within the first six months, flush with secondary sales, inventory and fulfillment data from the field, our client was able to recognize and plug gaps in fulfillment to the distributor. This in turn:
A US-based CPG firm needed to develop a real-time monitoring solution to predict stock outs, reduce breakdowns of vending machines and coolers with predictive analytics, and integrate these with image processing and social analytics to plug revenue leaks.
Leveraging Infosys IoT platform, the team retrofitted legacy vending machines enabling them with M2M for intelligence. Azure-based platforms for vending machine management and real-time analytics were designed and installed for devices. Infosys also analyzed the landscape to derive analytics–led insights, across geographies, by co-relating flavors, weather conditions and sales.
A global mining major was keen to have real-time visibility across the mining value chain – extraction, transportation and processing – to be able to manage operations effectively.
Leveraging Infosys IoT platform, we helped the firm enable integration and interoperability between their multiple manufacturing systems, historical data and legacy systems. Critical systems were moved to newer converters without disrupting business continuity thereby establishing a more stable platform for the existing quality system supporting smelter operations.