Automate
Everything

Software bots that replicate human interactions with digital systems to execute rule-based, repetitive tasks. Bots interact with UIs exactly as a human would — clicking, typing, reading screens — without requiring API integration. Deterministic, auditable, but brittle to UI changes.

RPA augmented with AI/ML capabilities — including NLP, computer vision (CV), OCR, and machine learning — to automate complex, judgment-intensive processes involving unstructured data. IPA handles what rules-based RPA alone cannot: invoices, emails, images, and exception-heavy workflows.

A broad category: using technology to automate recurring, multi-step business workflows. BPA encompasses RPA, workflow engines, BPM platforms, and API integrations. Focuses on end-to-end process flows rather than individual task automation. Governed by BPMN (Business Process Model & Notation) standards.

Gartner's 2019 term for "a business-driven, disciplined approach to rapidly identify, vet and automate as many business and IT processes as possible" using an orchestrated combination of RPA, AI, ML, BPM, iPaaS, process mining, and low-code tools. Enterprise-wide automation at scale — not isolated bot deployments.

RPA deployment where a bot runs on a human's workstation and is triggered by the user during their workflow. Human and bot collaborate in real time — the person handles judgment while the bot handles repetitive sub-tasks. Common in call centers, front-office operations, and compliance-heavy processes.

RPA deployment where bots execute autonomously on servers without human presence or triggering. Scheduled or event-driven. Processes entire queues of work end-to-end — invoices, reports, data migrations. Requires robust exception handling and monitoring since no human is present to intervene mid-process.

A virtual employee — a software entity (bot or AI agent) assigned a persona, skillset, and work queue. Digital workers can be assigned tasks, managed, and measured on output like human FTEs. The concept abstracts individual bots into workforce-level resource management. Used by UiPath, Automation Anywhere, and Blue Prism.

A data-driven technique extracting knowledge from event logs in enterprise systems (ERP, CRM, ITSM) to automatically reconstruct, visualize, and analyze actual business process flows. Objectively identifies deviations, bottlenecks, and automation candidates. Key vendors: Celonis, UiPath Process Mining, Microsoft Process Advisor.

The capture and analysis of user desktop interactions — keystrokes, mouse clicks, app navigation — to understand how humans perform work at the task level. Complements process mining (which uses system logs) by capturing the human execution layer. Generates task flow diagrams and surfaces RPA automation candidates.

Development platforms enabling automation creation with minimal or zero hand-written code, using visual drag-and-drop interfaces, pre-built connectors, and declarative logic. Low-code targets developers; no-code targets business users. Examples: Power Automate, Zapier, n8n, Make. A named component in Gartner's hyperautomation stack.

A software design paradigm where components communicate by producing and consuming events (messages representing state changes). Producers emit events to a broker (Kafka, RabbitMQ, AWS EventBridge); consumers subscribe and react asynchronously. Enables real-time, decoupled automation at enterprise scale. Explicitly named in Gartner's hyperautomation definition.

The coordination and sequencing of automated tasks, bots, agents, and services into a coherent end-to-end process. An orchestrator manages execution order, dependencies, error handling, retries, and monitoring. Examples: UiPath Orchestrator, Apache Airflow, Temporal, n8n, Azure Logic Apps. The "conductor" of the automation stack.

An LLM-powered system that perceives context, reasons about goals using chain-of-thought, selects tools or actions, executes them, observes results, and iterates — all within a loop until the goal is achieved. Fundamentally different from RPA: agents handle ambiguity and open-ended tasks; bots execute fixed, deterministic scripts.

An architecture where multiple specialized AI agents collaborate — an orchestrator agent delegates subtasks to specialist agents (researcher, coder, writer, critic), each with their own tools, memory, and instructions. Enables parallelism and specialization beyond a single agent. Frameworks: LangGraph, CrewAI, Microsoft AutoGen.

The capability of an LLM-based agent to invoke external functions — web search, code execution, database queries, API calls, file operations — to gather information or take actions beyond text generation. Tools are declared to the model; the model decides when and how to call them. The mechanism transforming LLMs from responders into actors.

A prompting and agent design framework (Reasoning + Acting) where an LLM agent alternates: Thought → Action → Observation, repeating until complete. The agent verbalizes reasoning before each action, then observes the result. Introduced by Yao et al. (arXiv:2210.03629, 2022). Reduces hallucinations by grounding reasoning in real-world observations.

An automation design pattern inserting human judgment, approval, or input at critical decision points within an otherwise automated workflow. Used when confidence is low, stakes are high, or compliance requires sign-off. Balances automation efficiency with oversight — especially critical in agentic AI deployments handling consequential actions.

A practice treating automation as an ongoing, iterative discipline rather than a one-time project — continuously discovering, building, deploying, monitoring, and optimizing automations as processes evolve. Mirrors DevOps principles: CI/CD pipelines for bots. Enabled by process mining feedback loops, bot health monitoring, and automated ROI dashboards.