The Enterprise Automation landscape is a fast-evolving space and new terminologies keep emerging – Hyperautomation, Low-Code, No-Code, AI, ML, NLP, NLG, Computer Vision, Process Mining, ICR/OCR, RPA, RDA, Assisted Bots, Unassisted Bots, Chatbots – the list is long. Here’s our attempt to demystify the lingo!
Artificial Intelligence (AI): The simulation of human intelligence processes by computer systems. In the context of automation, AI tools fall under a few broad categories: perception (the acquisition of unstructured information), cognition (rules for processing that information) and conversation (exchange of the information using an unstructured format).
Assisted Automation: RPA or Cognitive Bots that are triggered on a “as-needed” basis by human agents in order to help them complete their tasks more efficiently and/or accurately. These bots run on the same computer as the human agent and their typical uses cases are in the Front Office or Contact Centre of an enterprise. Assisted Automation is also known as Attended Automation or Remote Desktop Automation (RDA).
Attended Automation: See Assisted Automation.
Blockchain: A highly-secure, digital, distributed ledger technology that can be used for applications such as “smart contracts” within the enterprise. Typical use cases include financial transactions, legal and intellectual property protection.
Business Process Management (BPM): Methodologies and technologies that help orchestrate and streamline processes in the enterprise using a workflow-based approach. Enterprises should view BPM as a complimentary technology to intelligent automation using RPA, AI and ML tools.
Case Management: A solution approach typically used for recording, tracking and solving customer issues across sales, service and support. It is a tool to ensure creating a responsive organisation.
Chatbots: Chatbots or Conversational AI refers to the use of messaging apps and speech-based assistants to automate the “human-machine conversation” in order to create hyper-personalised customer experience at scale. These applications leverage AI tools like NLP and Deep Learning to enable long-running, free-flow interactions with customers using natural language (text or voice).
Cognitive Bots: RPA technology that combines the ability to execute rule-based, repetitive tasks with AI or ML capabilities.
Complex Event Processing: A solution for near real-time processing of streaming information (data) to drive business decision-making and responses. Typical examples include social media (twitter feeds), stock market feeds and IOT sensor data feeds.
Computer Vision: One of the application areas of Machine Learning that uses image processing algorithms to interpret images or videos to extract content and context (representing a type of unstructured data).
Conversational AI: See Chatbots.
Deep Learning (DL): A technique for implementing Machine Learning in which the algorithms use “neural networks” (inspired by the information processing patterns found in the human brain). While traditional ML algorithms require the features to be used for data classification to be fed to it, DL algorithms can automatically discover them.
Democratisation: A federated approach that makes technology and automation tools available to end users or “citizen” users for an accelerated digital transformation of the enterprise. Low-code, no-code platforms are one of the key enablers that can help democratisation in areas such as application development, analytics and data-driven decision making.
Digital Twin: An approach focused on creating a “replica” of the people-process-system interactions within the enterprise in order to visualize them better and identify opportunities for automation and driving business outcomes.
Emotion AI: Application area of AI focused on using tools such as NLP, text analytics and biometrics to identify and quantify the positive or negative sentiment in a customer interaction in order to provide an appropriate intervention.
Ethical AI: A responsible approach to leveraging Artificial Intelligence and Machine Learning within the enterprise based on corporate ethics and values. The focus is to build transparency and traceability into how these techniques are used, especially in decision making related to customer-facing interactions.
Intelligent Character Recognition (ICR): An extended and more-specific technology within OCR that can extract information from hand-written documents as against printed characters. Like OCR, this is a key technology area that complements the RPA and AI tools by providing them with the data to work with.
Intelligent Document Processing (IDP): A key lever for enabling Hyperautomation solutions, IDP focuses on intelligent extraction, classification, verification and processing of data from documents in formats such as PDF, JPG and PNG. This unstructured to structured data conversion uses techniques such as OCR/IC, ML/DL and NLP for extraction and comprehension of information.
Intent Analysis: Application area of AI focused on using NLP, text analytics and Deep Learning in order to determine propensity of a customer to engage and convert and achieve a desired business outcome. This technique is often used in conjunction with Emotion AI or Sentiment Analysis to drive prospects through a customer journey.
Hyperautomation: Hyperautomation is the advancement of process automation that leverages advanced technologies like Artificial Intelligence, Machine Learning, Intelligent Document Processing, Cognitive RPA, iBPMS, Process Discovery/Mining and “Human-in-the-loop” to create more impactful business outcomes.
Hyper Intelligent Automation: See Hyperautomation
Low-code, No-Code: A modern way of building enterprise applications with minimal coding, using a “drag-drop-and-configure” interfaces. A robust Low-code, No-code platform must have the ability to develop and run a diverse set of applications including – UI based applications (web. mobile, web), Data Processing, Event Processing, Process Automation, Blockchain, Internet-Of-Things (IOT).
Machine Learning (ML): ML is a specific area of AI that focuses on getting a computer to act without programming by looking for patterns. The broad categories of ML are supervised learning (where data sets are labelled so that patterns can be detected and used to label new data sets), unsupervised learning (where data sets are not labelled but instead sorted according to similarities and differences) and reinforced learning (where data sets are not labelled and learning happens by the feedback based on the actions performed).
Natural Language Generation (NLG): NLG is a subtopic of NLP focused on converting structured data into human understandable formats like text and voice.
Natural Language Processing (NLP): The processing of information in human (and not computer) language by a computer program. NLP use cases in the context of automation include voice-to-text conversion, text translation, sentiment analysis and intent analysis. NLP is the summation of NLU and NLG and is key to facilitating the “human-machine conversation”.
Natural Language Understanding (NLU): NLU is a subtopic of NLP focused on the machine’s comprehension capabilities of human inputs like text and voice and converting it into structured data. Most common examples of NLU are Siri, Alexa and Google Assistant.
Optical Character Recognition (OCR): Technology that can help recognise text inside images such as scanned documents and photos and convert it into machine-readable text data. OCR algorithms processes a digital image by locating and recognising characters (letters, numbers and symbols). OCR is an important complementary technology that has helped drive the recent success of RPA and AI.
Process Discovery: An automation enabler technology based on recording of human interactions with enterprise systems, productivity tools and external interfaces to identify new automation opportunities. Process Discovery follows a non-intrusive approach for analysing the end-to-end business process including human-to-system and system-to-system interactions.
Process Mining: A technology tool focused on driving process optimisation by reducing the deviation from the “happy path” of the business process, leading to effort and cost efficiencies. Process Mining uses an approach of running pattern analysis on system logs generated by enterprise systems such as ERP and CRM applications.
Robotic Desktop Automation (RDA): See Assisted Automation.
Robotic Process Automation (RPA): See Unassisted Automation. RPA also represents a broader definition of representing the automation technologies in which tasks are carried out by a computer by mimicking human behaviour or “taking the robot out of the human”.
Sentiment Analysis: See Emotion AI.
Structured Data: Refers to information that is highly organised such as that in a relational database, thereby making it easy for machines to process.
Unassisted Automation: RPA or Cognitive Bots that are run in batch mode to carry our high-volume, repetitive tasks without having to be triggered by a human agent. Typical use cases for Unassisted Automation are in the Back Office. These bots typically run on a server and not on the same computer as a human agent.
Unattended Automation: See Unassisted Automation.
Unstructured Data: Unstructured data is information that is not specifically structured to be easily understood by machines. The most common format of unstructured data is text, but it could have other formats including images, video, audio or social media data.