What is it?
Blockchain is essentially a de-centralized ledger concept that contains transactional entries that must be confirmed and encrypted. There is no central hub where transactional data is stored but rather servers and hard drives all over the world that hold blocks of data for additional security. This provides a way to securely and effectively create a protected log of sensitive activity. It can reduce fraud, increase financial privacy, and speed up transactions.
Examples of how customers use the technology in business
For real-time transactions that span multiple parties, ERP always had limitations on data transparency. Through a concept of UUID (unique universal identifier), our customers in aerospace and process industries are using this to create a shared ledger that can be shared by multiple parties (business partners, customers, vendors). This has created completely new interactions (some even replacing RFID, Bar-code and EDI transactions), to share records instantaneously on the block chain network, and enable microtransactions that previously were not viable and secure.
What is it?
Internet connectivity of a network of physical objects that are embedded with software/sensors to allow the collection and exchange of data. Includes objects like wearables (watches), buildings, appliances, and vehicles).
Examples of how customers use the technology in business
In the space of IoT, our solutions encompass hardware (beacons, sensors, readers, access points, gateways, etc.) and software (mostly API and cloud storage, with ability to do analytics and trending) to serve several business applications to track in real-time fixed and mobile assets, with real-time location requirements, using BLE (blue-tooth), WiFi and Radio frequencies to transmit data with or without battery requirements in some equipment used.
Our customer use cases include:
- Manufacturing – Tracking pallets, work in progress, finished goods, improving yield, traceability
- Distribution – asset tracking in distribution centers, warehouses and vehicles, cold chain, delivery verification, counterfeit and grey market detection
- Retail – Consideration measurement, advertising attribution, triggering media phones and video displays
- Ownership – automatic replenishments, Product information and instructions, product finding, wardrobe management, and recommendation
What is it?
Analysis of extremely large data sets that can be high in volume, velocity or variety.
Examples of how customers use the technology in business
We have used the public clouds as well as commodity hardware on-premise to build a Hadoop cluster for data storage and used open-source software from Apache to help store primarily unstructured data and then allow its link to structured data in various relational database.
Use cases include: Storing images and doing predictive analysis – for example, we had linked image color streaming patterns on ocean bed assets and asset maintenance data for doing preventative maintenance and compliance reporting. We have also used image recognition to help customers gain insight from images (logo, label detection, etc.), Determine inappropriate content. Other use cases include social data capture and intelligent mining for specific search patterns (warning/alerting) and sentiment analysis.
What is it?
AI is the simulation of human intelligence processes by machines/computers. Examples include speech recognition, decision-making and visual perception. Machine Learning is a subset off artificial intelligence where systems use algorithms and statistical models to identify patters in data, learn from it and perform a task without explicit programming and only minimal human intervention.
Examples of how customers use the technology in business
Areas where we are working on proof of concepts are depicted in the infographic below.
Automation or Artificial intelligence started a few years back with us partnering with Canon to doing optical character recognition and doing AP automation – over past several years, we have worked with several customers automating AP functions and having human intervention only to exception handling. Today we are extending the use of AI into machine learning using Google TensorFlow, Keras, Apache Spark/ MXNET, Facebook PyTorch, Berkley’s CAFFE, Sonnett, Microsoft’s CNTK and many more platforms to serve customers use statistical algorithms in space of statistical (regression, classification, segmentation clustering – supervised, unsupervised and reinforced learning) to develop AIML solutions. Use cases include RPA (robotic process automation – AP automation was one of the early ones, but we have several in space of Customer order processing, website scraping, customer email replies, payroll and forms processing, etc.), Fraud detection (combining credit card checks with human recognition techniques and avoiding machine spoofing of our customers website), product recommendations on our website, campaign effectiveness and demand forecasting, plant maintenance and monitoring. We use several algorithmic techniques in conjunction with platforms using R, Python and other programming language integrations to deploy techniques such as Naiive Bayes, Monte Carlo, K-means, Support Vector machines, Apriori (filtering), decision trees, ensemble learning (both random forests and gradient boosted decision trees) and neural networks (with optimizer and deep learning techniques), where we use spot cloud instances to run large loads (datasets) and numbers for customers, that were impossible in days and years gone by and economically non-viable, which are very much in reach and feasible today.