Enterprise AI and Process Intelligence Glossary

The language of enterprise AI, process intelligence, and intelligent automation is evolving fast. This glossary defines the key terms used across Skan AI's platform, content, and industry conversations, so you can follow the discussion without needing a decoder. Whether you are new to process intelligence or evaluating where agentic AI fits your organization, start here.

Glossary

Demystifying The Terms Driving Business Transformation

A

  • Accounts Payable (AP)
    Accounts Payable (AP) is the business function managing payment obligations to suppliers, covering invoice processing, approvals, and payment execution. Skan AI observes how AP teams handle invoices, identifying the manual touchpoints and exception patterns with the greatest automation potential.

  • Action
    An action is the smallest observable unit of work captured by Skan AI, such as opening an application, entering data, or clicking a field. Skan's computer vision classifies actions automatically, transforming raw desktop activity into structured process data that makes previously invisible work measurable.

  • Activity
    An activity is a group of related actions that together complete a defined step in a business process. Skan AI identifies activities automatically from desktop observation, replacing manual process interviews with real-time data on how work actually happens.

  • Agency Gap
    The agency gap is the divide between knowing what to do and being able to act on it autonomously. Gartner coined this to describe the limitation of GenAI: it informs but does not act. Skan AI closes the agency gap for enterprise processes by grounding agents in operational ground truth.

  • Agent
    An AI agent is an autonomous software system that perceives its environment, makes decisions, and takes actions to reach a goal without step-by-step human instruction. Skan AI builds agents from operational ground truth, real observed work data rather than assumed process maps, so agents perform accurately from day one.

  • Agent Memory / Persistent Context
    Agent memory is the capability of an AI agent to retain and build on information from previous interactions and tasks, improving performance over time. Skan AI's context graph of work provides persistent process memory, ensuring each agent task is informed by the full history of how that process has been observed to behave.

  • Agent Operating Procedures
    Agent Operating Procedures (AOPs) are structured, machine-readable process definitions governing how AI agents execute tasks, including workflow steps, decision rules, escalation thresholds, and success criteria. Skan AI generates AOPs from observed work data, ensuring agents receive accurate operational playbooks rather than idealized documentation.

  • Agent Washing
    Agent washing is the practice of rebranding existing automation tools as AI agents without implementing genuine agentic reasoning. Gartner coined this to warn buyers about inflated vendor claims. Skan AI recommends evaluating any claimed agent on whether it reasons from context, adapts to change, and is grounded in real process data.

  • Agent-to-Agent Protocol (A2A)
    The Agent-to-Agent Protocol (A2A) is an open standard enabling AI agents from different platforms to share context and coordinate tasks securely. Gartner pairs A2A with MCP as foundational enterprise agentic architecture standards.

  • Agentic AI
    An autonomous form of artificial intelligence that independently makes decisions, learns in real-time, and adapts to dynamic environments with minimal human intervention.

B

  • Bottleneck Analysis
    The process of identifying stages in a workflow or system that slow down operations or create inefficiencies.

  • BPMN (Business Process Model & Notation)
    BPMN is the standardized visual language for modeling enterprise processes.

  • Business Process Automation
    Business process automation uses technology to execute recurring business processes with minimal human intervention.

  • Business Process Intelligence
    Business process intelligence is the capability to analyze and generate insights from enterprise process data.

C

  • Center of Excellence (CoE)
    A team that governs and scales an organization's automation and AI initiatives.

  • Compliance
    Process compliance is verification that business processes are executed in line with regulatory requirements and internal policies.

  • Computer Vision
    Computer vision is an AI technology that interprets visual information, including screen content.

  • Confidence Threshold
    A confidence threshold is the minimum certainty level an AI agent must reach before acting autonomously.

D

  • Data Discovery
    Process data discovery is the capture and analysis of data generated by how work is actually performed.

  • Days Payable Outstanding (DPO)
    Days Payable Outstanding (DPO) measures the average days an organization takes to pay suppliers after receiving goods or services.

E

  • Enterprise AI
    Enterprise AI refers to the application of artificial intelligence technologies within organizational settings.

  • Event Log
    An event log is a structured record of timestamps and activities captured by enterprise systems.

F

  • First-Party Process Data
    Work activity data captured directly from an organization's operations through real-time desktop observation.

G

  • Generative AI
    Generative AI is a subset of artificial intelligence that focuses on creating new content or data.

H

  • Human-in-the-Loop (HITL)
    Human-in-the-loop (HITL) is the design principle of embedding specific human review points within automated workflows.

I

  • Intelligent Automation
    Intelligent automation combines AI with process automation to handle judgment-intensive tasks beyond the reach of traditional RPA.

  • Intelligent Document Processing (IDP)
    Intelligent Document Processing (IDP) extracts, classifies, and validates information from unstructured documents.

J

  • Job Evaluation
    A systematic process of determining the relative worth of jobs within an organization.

L

  • Large Action Models
    Large Action Models (LAMs) are AI foundation models trained on human application interaction data.

M

  • Manual Process Discovery
    Manual process discovery maps business processes through workshops, interviews, and manual observation.

N

  • Neurosymbolic AI
    Neurosymbolic AI combines neural network pattern recognition with symbolic rule-based reasoning.

O

  • Object-Centric Process Mining
    Object-centric process mining analyzes how multiple business objects interact within a single process.

P

  • Predictive Analytics
    Predictive analytics uses data to forecast future process performance.

  • Process Automation
    Process automation involves using technology to execute recurring tasks or processes in a business.

R

  • Robotic Process Automation (RPA)
    Robotic Process Automation (RPA) uses software bots to execute repetitive tasks by mimicking human interactions.

S

  • Skan Agents
    Skan Agents is the agentic execution layer of the Skan AI platform.

  • Structured Observation Data
    Structured observation data is work activity data captured through real-time desktop observation.

T

  • Task Mining
    Task mining captures user interactions with software applications to analyze and identify patterns in task execution.

U

  • Unit Cost Analysis
    Unit cost analysis determines the true cost of completing one unit of work within a business process.

V

  • Virtual Assistant
    Skan's desktop observation agent is a software component that uses computer vision to observe workflow activity in real time.

W

  • Work Telemetry
    Work telemetry is the continuous capture of data signals from how employees interact with applications and workflows.