Skip to main content

AI 101: the words, without the fog.

Learning AI, or nodding along in meetings? Plain answers to the terms that never get explained: what each one means, and why it costs or saves you money. Free, no sign-up, and every term has its own page.
A strategist smiling in front of a whiteboard covered in planning notes

Twelve questions, in the order we would explain them.

About fifteen minutes end to end. Each one is its own short page with a one-sentence answer, why it matters to a budget, and what to read next.
  1. Start

  2. Then

  3. Before you spend

Every term, searchable.

41 terms. Search matches names, acronyms and definitions.

New to all of this?

Five terms carry most of the weight. Read these and the rest of the page makes sense.

The basics

Artificial intelligence

Also called AI

Software that produces useful output from messy input without being given explicit rules for every case. In practice, when someone says AI today they almost always mean a large language model.

Why it matters. The word covers everything from a spam filter to a chatbot, so it is nearly useless in a procurement conversation. Ask which specific technique is being proposed and what it does when it is wrong.

What is artificial intelligence? Read the full answer

See also LLM (large language model), Deterministic logic

LLM (large language model)

Also called LLM, large language model, foundation model, GPT

A model trained on very large amounts of text that predicts what text should come next. Claude, GPT and Gemini are all large language models.

Why it matters. It predicts plausible text. It does not look anything up unless you connect it to something. That single distinction explains most AI project failures.

  • Claude
  • GPT
  • Gemini
  • Llama
  • Mistral

What is a large language model (LLM)? Read the full answer

See also Hallucination, RAG (retrieval augmented generation)

Prompt

The instruction you give a model. In a production system the prompt is usually assembled by code rather than typed by a person.

Why it matters. Prompts are code. They need version control, review and tests. Treating them as copy that anyone can tweak is how a working system quietly stops working.

What is a prompt? Read the full answer

See also Context window, Evaluation

Token

The unit a model reads and writes. Roughly three quarters of a word in English. Providers bill per token, in and out.

Why it matters. Token count is your unit cost. A feature that looks cheap in a demo can be expensive at volume, and the difference is usually how much context gets sent on every single request.

What is a token in AI? Read the full answer

See also Context window, Inference

Context window

How much text a model can consider at once, prompt and response together. Measured in tokens.

Why it matters. Everything the model knows about your request has to fit in here. It is why systems get built to fetch the right ten paragraphs rather than send the whole handbook every time.

What is a context window? Read the full answer

See also Token, RAG (retrieval augmented generation)

Inference

One run of a model. Sending a request and getting an answer back.

Why it matters. Training is a one-off cost that someone else usually paid. Inference is the recurring one you pay per use, forever. Model choice is largely an inference cost decision.

What is inference in AI? Read the full answer

See also Token

Deterministic logic

Ordinary code. Same input, same output, every time, and you can read the rule that produced it.

Why it matters. A great deal of what gets proposed as AI is better done as a rule. Deterministic logic is cheaper, faster, testable and auditable. Reach for AI where the input is genuinely unpredictable, not because it is on the roadmap.

When deterministic logic beats AI

What is deterministic logic, and when does it beat AI? Read the full answer

See also Artificial intelligence

Generative AI

Also called GenAI, gen AI

AI that produces new content, such as text, images, code, audio or video, in response to a request. ChatGPT, Claude and Gemini are generative AI products built on large language models.

Why it matters. It is the kind of AI most budgets are being spent on today. Knowing that it generates a likely answer rather than looking one up explains both what it is good at and why it needs checking.

  • ChatGPT
  • Claude
  • Gemini
  • Midjourney

What is Generative AI? Read the full answer

See also LLM (large language model), Hallucination, Agentic AI

Prompt engineering

Also called prompting, system prompt, context engineering

The work of writing and structuring the instructions and background given to a model so it produces a useful, consistent result. It covers the system prompt, examples, and which information is supplied with each request.

Why it matters. It is the cheapest lever you have. A better prompt often fixes a quality problem that teams try to solve by buying a larger model, and a shorter one cuts the bill.

What is Prompt engineering? Read the full answer

See also Prompt, Context window, Token

How AI systems are built

AI agent

Also called agent, AI agents, autonomous agent

A model given tools and a goal, allowed to decide which tools to use and in what order, looping until it thinks it is finished.

Why it matters. The step change is that it acts rather than answers. That is also the risk: an agent with write access can do real damage quickly, so the interesting questions are what it can reach and who approves what.

  • Claude Code
  • Cursor
  • GitHub Copilot Agent

What the MCP release changes for enterprise agents

What is an AI agent? Read the full answer

See also MCP (Model Context Protocol), Human in the loop, Orchestration

MCP (Model Context Protocol)

Also called MCP, Model Context Protocol, MCP server

An open standard for connecting AI models to tools and data sources. One protocol instead of a bespoke integration per model and per system.

Why it matters. Before it, every AI integration was custom, so switching model provider meant rebuilding the plumbing. MCP is what makes vendor neutrality practical rather than aspirational.

  • Claude Desktop
  • Cursor
  • VS Code

Stateless, secure and extendable

What is the Model Context Protocol (MCP)? Read the full answer

See also AI agent, API

RAG (retrieval augmented generation)

Also called RAG, retrieval augmented generation, grounding

Fetch the relevant documents first, put them in the prompt, then ask the model to answer using them.

Why it matters. It is the cheapest way to make a model answer from your data rather than its training. Try this before anyone proposes fine-tuning or a custom model.

  • Azure AI Search
  • LlamaIndex
  • LangChain

What is retrieval augmented generation (RAG)? Read the full answer

See also Fine-tuning, Vector database, Hallucination

Vector database

Also called embeddings database

A store that finds text by meaning rather than by keyword, by comparing numerical representations called embeddings.

Why it matters. It is the retrieval half of retrieval augmented generation. Worth knowing it is a component you can swap, not a platform you marry.

  • Pinecone
  • Weaviate
  • pgvector
  • Qdrant

What is a vector database? Read the full answer

See also RAG (retrieval augmented generation), Embedding

Embedding

A list of numbers representing a piece of text, arranged so that similar meanings sit near each other.

Why it matters. It is what lets a search find the right answer when the user typed none of the same words the document uses.

What is an embedding? Read the full answer

See also Vector database

Fine-tuning

Continuing to train an existing model on your own examples so it adopts a particular style, format or task.

Why it matters. Expensive, slow to change, and it does not teach the model facts reliably. It is the right answer far less often than it is proposed. Try retrieval first.

What is fine-tuning? Read the full answer

See also RAG (retrieval augmented generation), Inference

Orchestration

The code that decides what happens in what order: which model, which tool, what to do on failure, when to ask a person.

Why it matters. This is where most of the actual engineering lives. The model is a component. The orchestration is the system, and it is what you own.

  • LangGraph
  • Temporal
  • n8n

What is AI orchestration? Read the full answer

See also AI agent, Human in the loop

API

A defined way for one system to ask another for something. The contract between two pieces of software.

Why it matters. If a platform has no usable API, nothing else you want to do is possible. It is the first thing to check, before features.

What is an API? Read the full answer

See also MCP (Model Context Protocol), Microservices

Agentic AI

Also called agentic, agentic workflows, agentic commerce

AI that takes actions toward a goal instead of only answering a question. An agentic system plans steps, uses tools such as search, a CRM or a checkout, checks the result and carries on until the task is done.

Why it matters. It is where AI moves from saving minutes to doing work. It is also where permissions, audit trails and cost controls start to matter, because the system is now acting on your behalf.

  • A shopping assistant that builds a basket
  • An agent that triages a sales inbox

AI & Automation

What is Agentic AI? Read the full answer

See also AI agent, Tool calling, Multi-agent system, Human in the loop

Tool calling

Also called function calling, tool use, actions

The mechanism that lets a model ask your software to do something: look up an order, search a catalog, send an email. The model decides which tool to use and with what inputs, and your code runs it.

Why it matters. It is the difference between an assistant that talks about your business and one that can work in it. Every tool is also a permission you are granting, so the list should be short and deliberate.

What is Tool calling? Read the full answer

See also AI agent, MCP (Model Context Protocol), API, Guardrails

Multi-agent system

Also called multi-agent, agent swarm, agent orchestration

Several AI agents, each with a narrow job, working together on a larger task. One might research, one write and one check, with a coordinator passing work between them.

Why it matters. Narrow agents are easier to test and cheaper to run than one agent asked to do everything. The cost is coordination, so it only pays when the task really has separate parts.

  • LangGraph
  • CrewAI
  • AutoGen

What is Multi-agent system? Read the full answer

See also AI agent, Orchestration, A2A (agent-to-agent)

A2A (agent-to-agent)

Also called A2A, agent2agent, agent interoperability

An open protocol that lets AI agents built by different companies find each other, exchange tasks and report results. MCP connects an agent to tools. A2A connects an agent to other agents.

Why it matters. Your customers' agents will soon talk to your agents: a shopping assistant asking your pricing agent for a quote, for example. A shared protocol is what stops that needing a custom integration per partner.

What is A2A (agent-to-agent)? Read the full answer

See also MCP (Model Context Protocol), Multi-agent system, Agent experience

Open-source model

Also called open-weight model, open model, self-hosted LLM, local LLM

A model whose weights are published, so you can run it on your own hardware or cloud account instead of calling a vendor's API. Llama, Mistral and Gemma are common examples.

Why it matters. It gives you control over where data goes and can cut running costs at high volume. You take on the hosting, the updates and the security yourself, so it is a trade and not a free option.

  • Llama
  • Mistral
  • Gemma
  • Ollama

What is Open-source model? Read the full answer

See also LLM (large language model), Inference, Data residency

Making AI reliable

Hallucination

A model producing confident text that is wrong. Not a bug that gets fixed, a property of how the technology works.

Why it matters. Any design that assumes correct output is already broken. Systems that survive contact with production ground answers in retrieved sources and put a person in front of anything expensive to get wrong.

How we handle it

What is an AI hallucination? Read the full answer

See also RAG (retrieval augmented generation), Human in the loop, Evaluation

Evaluation

Also called evals

A written set of test cases and the criteria for a correct answer, run against the system so you can tell whether a change made it better or worse.

Why it matters. This is the single most skipped step and the most common reason a pilot never ships. Without it nobody can prove the new version is an improvement, so the team keeps swapping models and hoping.

  • Braintrust
  • LangSmith
  • Promptfoo

Your pilot is not failing on the model

What is AI evaluation? Read the full answer

See also Hallucination, Prompt

Guardrails

Checks around a model that block or alter unacceptable input and output, separate from the model itself.

Why it matters. Instructing a model to behave is a request. A guardrail is a control. Compliance conversations need the second one.

  • Llama Guard
  • Azure Content Safety
  • NeMo Guardrails

What are AI guardrails? Read the full answer

See also Human in the loop, AI governance

Human in the loop

A person reviews or approves before an action takes effect, rather than after.

Why it matters. It is both the practical answer to hallucination and, for high-risk uses, a legal requirement under the EU AI Act. Design where it sits early; retrofitting approval into an automated flow is painful.

Check your AI Act risk tier

What does human in the loop mean? Read the full answer

See also High-risk AI, Guardrails

Shadow AI

AI tools used inside the organization that nobody approved, procured or knows about.

Why it matters. It is where your data actually leaks. One person installing a tool and pointing it at a repository is enough, and no policy document prevents it on its own.

One person on your IT team used Grok

What is shadow AI? Read the full answer

See also Data residency, AI governance

AI cost optimization

Also called reduce AI costs, LLM cost optimization, AI spend, FinOps for AI

The practice of lowering what AI costs to run without lowering what it delivers. The main levers are sending fewer tokens, using a smaller model where it is good enough, caching repeated requests, and replacing AI with plain rules where rules do the job.

Why it matters. AI is priced by usage, so success raises the bill. Teams that measure cost per task from the start can cut spend substantially. Teams that do not find out when finance asks.

Reduce AI costs

What is AI cost optimization? Read the full answer

See also Token, Inference, Deterministic logic, Evaluation

Rules and responsibility

EU AI Act

The European Union's regulation on artificial intelligence. It sorts AI uses into risk tiers and attaches obligations to each, with fines up to 35 million euros or 7 percent of worldwide turnover.

Why it matters. It applies based on where the system is used, not where your company is. The prohibitions and the transparency duties are already in force. The high-risk obligations follow.

  • Regulation (EU) 2024/1689

Check your risk tier in 5 minutes

What is the EU AI Act? Read the full answer

See also High-risk AI, Provider vs deployer, AI governance

High-risk AI

Under the EU AI Act, AI used in areas like hiring, credit, education, essential services or biometrics. It carries the heaviest obligations: risk management, documentation, human oversight and a conformity assessment before go-live.

Why it matters. It catches more organizations than expected. If AI touches a decision that materially affects a person, assume this tier until you have checked.

Find out which tier you are in

What counts as high-risk AI under the EU AI Act? Read the full answer

See also EU AI Act, Human in the loop

Provider vs deployer

Under the EU AI Act, a provider builds and puts an AI system on the market. A deployer uses one. The obligations differ sharply.

Why it matters. Most organizations assume they are only deployers and are wrong, because building on top of a model can make you a provider. You also cannot rely on a vendor having done its part: buying it is not a defense.

Work out which you are

Provider vs deployer: who is responsible under the EU AI Act? Read the full answer

See also EU AI Act, High-risk AI

AI governance

Knowing which AI systems you have, what each is used for, who owns it, and being able to show that on request.

Why it matters. It sounds like paperwork and is mostly an inventory problem. Organizations that cannot list their AI systems cannot comply with anything, and that is the usual starting position.

What is AI governance? Read the full answer

See also EU AI Act, Guardrails

Data residency

Which country or region your data is physically stored and processed in.

Why it matters. It constrains model and hosting choice, so it is a design input rather than a deployment detail. Discovering it late usually means rebuilding.

What is data residency? Read the full answer

See also Shadow AI, AI governance

Composable architecture

MACH

Microservices, API-first, Cloud-native, Headless. A set of principles for building systems out of replaceable parts rather than one large platform.

Why it matters. The point is being able to change one thing without a replatform. That is what a monolith takes away and what it costs you every time the market moves.

  • commercetools
  • Contentstack
  • Netlify
  • Algolia

Score your composable maturity(opens in a new tab)

What is MACH architecture? Read the full answer

See also Composable architecture, Headless, Microservices

Composable architecture

Assembling a system from best-fit components connected by APIs, so any one can be replaced without rebuilding the rest.

Why it matters. It buys optionality, and it costs integration work. Done without senior architecture it produces a distributed monolith, which is worse than what it replaced.

When to go composable and when to stay put

What is composable architecture? Read the full answer

See also MACH, Microservices

Headless

A system that manages content or commerce and serves it over an API, with no opinion about how it is displayed.

Why it matters. It is what lets one source of content feed a website, an app, a kiosk and now an AI agent. It also means the front end is now your responsibility.

  • Contentful
  • Storyblok
  • Contentstack
  • Sanity

What is a headless CMS? Read the full answer

See also MACH, Digital experience platform

Microservices

Building an application as a set of small independent services that talk over APIs, rather than one deployable unit.

Why it matters. Independent deployment is the benefit. Operational complexity is the price. Both are real, and small teams routinely underestimate the second.

  • commercetools
  • Algolia
  • Stripe

What are microservices? Read the full answer

See also MACH, API

Digital experience platform

Also called DXP

A suite that bundles content management, personalization, analytics and campaign tools into one product.

Why it matters. Convenient at the start, expensive to leave. The composable question is whether you want one vendor's version of every capability or the best available version of each.

  • Sitecore
  • Adobe Experience Manager
  • Optimizely

CMS, DXP and MACH compared

What is a digital experience platform (DXP)? Read the full answer

See also Headless, Composable architecture

Being found by AI

AEO and GEO (answer engine optimization)

Also called AEO, GEO, generative engine optimization, AI search optimization, LLM SEO

Making your content easy for AI assistants to find, understand and quote, rather than optimizing for a ranked list of blue links.

Why it matters. People increasingly ask an assistant instead of searching. If the assistant cannot read your site, you are not in the answer, and no amount of traditional ranking fixes that.

  • ChatGPT
  • Perplexity
  • Claude
  • Google AI Overviews

AI is answering your customers

What is answer engine optimization (AEO)? Read the full answer

See also Agent experience, Structured data

Agent experience

Also called AX

How well a site works for an AI agent trying to accomplish something on it, as opposed to how it looks to a person.

Why it matters. Agents are becoming a real share of traffic, and they fail on things a person tolerates: content behind script, no machine-readable pricing, forms that need a mouse.

  • CrawlIndex
  • llms.txt

See how the web treats AI agents(opens in a new tab)

What is agent experience (AX)? Read the full answer

See also AEO and GEO (answer engine optimization), Structured data

Structured data

Also called schema markup, JSON-LD

Machine-readable labeling embedded in a page that states what things are: this is a price, this is an author, this is a definition.

Why it matters. It removes guesswork for search engines and AI assistants. Cheap to add, and it is the difference between being quoted accurately and being paraphrased wrongly.

  • schema.org
  • JSON-LD

What is structured data? Read the full answer

See also AEO and GEO (answer engine optimization), Agent experience

Commonly confused

The pairs people actually search as "X vs Y", side by side.

  • Agent vs LLM

    AI agent

    A model given tools and a goal, allowed to decide which tools to use and in what order, looping until it thinks it is finished.

    LLM (large language model)

    A model trained on very large amounts of text that predicts what text should come next. Claude, GPT and Gemini are all large language models.

    The difference. A language model answers. An agent acts. The agent is the model plus tools, a goal, and permission to keep going until it decides it is finished. All the interesting risk lives in that last part.

  • RAG vs fine-tuning

    RAG (retrieval augmented generation)

    Fetch the relevant documents first, put them in the prompt, then ask the model to answer using them.

    Fine-tuning

    Continuing to train an existing model on your own examples so it adopts a particular style, format or task.

    The difference. Retrieval hands the model the right documents at question time. Fine-tuning changes the model itself. Retrieval is cheaper, updates instantly when your content changes, and is the right answer far more often than it gets chosen.

  • Provider vs deployer, and does high risk change it

    Provider vs deployer

    Under the EU AI Act, a provider builds and puts an AI system on the market. A deployer uses one. The obligations differ sharply.

    High-risk AI

    Under the EU AI Act, AI used in areas like hiring, credit, education, essential services or biometrics. It carries the heaviest obligations: risk management, documentation, human oversight and a conformity assessment before go-live.

    The difference. Provider and deployer is about who you are. High risk is about what the system does. You can be a deployer of a high-risk system and still carry serious obligations, and buying it from someone else is not a defense.

  • Composable vs MACH

    Composable architecture

    Assembling a system from best-fit components connected by APIs, so any one can be replaced without rebuilding the rest.

    MACH

    Microservices, API-first, Cloud-native, Headless. A set of principles for building systems out of replaceable parts rather than one large platform.

    The difference. MACH is a specific set of four principles: microservices, API-first, cloud-native, headless. Composable is the broader idea of assembling replaceable parts. Every MACH system is composable. Not every composable system is MACH.

  • AI vs plain code

    Artificial intelligence

    Software that produces useful output from messy input without being given explicit rules for every case. In practice, when someone says AI today they almost always mean a large language model.

    Deterministic logic

    Ordinary code. Same input, same output, every time, and you can read the rule that produced it.

    The difference. Deterministic logic gives the same answer every time and you can read the rule that produced it. AI handles input you could not write a rule for. Most things proposed as AI are better as a rule, and the rule is cheaper to run and possible to audit.

Every question, one page each.

Want to check where you actually stand?

Two of these terms have a free tool behind them. Both give you a result before asking for anything.

Great questions from inquiring minds

How to use this glossary, and how to learn AI without the hype.

I am new to AI. Where should I start?

Read the four terms in the Start row, in order: artificial intelligence, large language model, prompt, and token. Each takes about a minute, and the rest of the glossary makes sense after them.

What is the difference between AI, an LLM, and an AI agent?

AI is the broad field. A large language model is one kind of AI that predicts text. An agent is a model given tools, a goal, and permission to keep acting until it decides it is done. Most of the risk, and most of the value, lives in that last part.

Why does every definition include why it matters?

Because a definition alone does not help anyone decide anything. Each term says what it is and then why it costs or saves money, which is the half most glossaries skip.

Can I cite these definitions?

Yes. Every term has its own page at fidgetlabs.io/ai-101/ followed by the term, with structured data so search engines and AI assistants can quote and attribute it accurately.

Is there a version for my leadership team?

Yes. AI Foundations for Leadership is a half-day session covering what AI can actually do, what it costs to run, and what to ask a vendor. AI 101 is the free, self-serve version.

Still not sure which of these you need?

That is a normal place to be. Book a call and we will tell you which of this applies to your situation, and which is noise.