AI Learning · 08 / 08

Glossary

Every acronym used in the AI Learning section, with its full name and what it means. Plurals take an s.

A

ACL (Access Control List)
A list of rules that says which users or groups may read a document or record. Search results are filtered with it so people only see what they are allowed to see.
ADR (Architecture Decision Record)
A short, dated document that records one architecture decision together with its context, consequences and rejected alternatives.
AI (Artificial Intelligence)
Software that performs tasks which normally need human judgement, such as reading documents, classifying them and drafting text.
AI-DLC (AI-Driven Development Lifecycle)
The delivery method used to build this example. AI drafts and checks work, while people make the decisions and sign each gate.
AP (Accounts Payable)
The finance function that records what the company owes its suppliers and pays those bills.
API (Application Programming Interface)
A defined set of requests and responses that lets one software system call another.

B

BM25 (Best Matching 25 ranking)
A keyword-search ranking formula that scores each document by how well its words match the words in the query.

D

DAG (Directed Acyclic Graph)
A set of steps with dependencies that flow in one direction and never loop back. Here it runs the invoice pipeline.

E

ERP (Enterprise Resource Planning)
Business software that keeps finance, purchasing, inventory and other core records in one system.

G

GL (General Ledger)
The master record of every financial account in a company's books.
GRN (Goods Receipt Note)
A record that goods were physically received, together with the quantities accepted.

I

ID (identifier)
A unique reference code that identifies one record, such as a document, a vendor or a payment.

J

JSON (JavaScript Object Notation)
A plain-text format for structured data made of named fields and their values.
JSON-RPC (JavaScript Object Notation Remote Procedure Call)
A message format built on JavaScript Object Notation for calling functions on another system. The Model Context Protocol uses it for tool calls.

L

LLM (Large Language Model)
A neural network trained on very large amounts of text. It generates and interprets language.

M

MCP (Model Context Protocol)
An open standard that lets AI applications discover and call tools and data sources in one consistent way.

O

OCR (Optical Character Recognition)
Software that turns images of printed or scanned text into text a computer can search and process.
OIDC (OpenID Connect)
A sign-in standard built on OAuth 2.0 that proves who a user or a service is.

P

PDF (Portable Document Format)
A fixed-layout document file format, widely used for invoices and contracts.
PII (Personally Identifiable Information)
Data that can identify a person. It must be limited to what is needed and protected.
PO (Purchase Order)
A buyer's formal order to a supplier, stating the items, quantities and agreed prices.

R

RAG (Retrieval-Augmented Generation)
A method that first retrieves the relevant documents and then has a language model answer from that material only, with citations.
RPC (Remote Procedure Call)
Calling a function that runs on another machine as if it were local.

S

SDK (Software Development Kit)
A library of tools and code that makes a vendor's service easier to call from an application.
SES (Service Entry Sheet)
A record that a service was performed and accepted. It plays the role of a goods receipt for services.
SFTP (SSH File Transfer Protocol)
A way to transfer files securely over an encrypted Secure Shell connection.
SHA-256 (Secure Hash Algorithm, 256-bit)
A fingerprinting function that gives a file a fixed-length digest. Any change to the file produces a different digest.
SLA (Service-Level Agreement)
A committed target for how quickly or reliably a service will be delivered.
SSH (Secure Shell)
An encrypted protocol for logging in to, and transferring files with, a remote computer.
SSO (Single Sign-On)
Signing in once with a central identity provider and then reaching many systems without signing in again.

Build this for your business

This example uses a generic invoicing system, but the same pattern — orchestrated steps, governed tools, grounded knowledge and a documented lifecycle — fits most back-office processes. We can scope a pilot on your own documents and systems.

Book a discovery call