What does
that actually mean?
AI vocabulary, explained through things you might actually do. Start with the term you keep hearing.
AI basics · Claude and its tools · Content and social media · Client work
AI basics
Personal AI assistant →
A personal AI assistant is software that uses AI to help with tasks using information about you and your work.
It might organise notes, prepare a daily brief or draft a reply.
What it can actually do depends on the files, tools and permissions you give it.
AI workflow →
An AI workflow is a repeatable sequence of steps that uses AI for part of a task.
It connects an input, such as meeting notes, to an output, such as a client follow-up.
The workflow also defines what happens next: a person checks it, another tool receives it, or the process stops.
AI agent →
An AI agent is an AI model that works toward a goal by choosing its own next steps.
It reads, uses tools, checks what happened and decides what to do next.
A chatbot answers your message and stops. An agent keeps going until the job is done, or until it gets stuck.
Context engineering →
Context engineering means choosing and maintaining the information an AI system has available while it works.
That includes instructions, relevant documents, examples, conversation history and tool results.
The aim is not to give it everything. It is to give it the right information for the job.
Context window →
A context window is the amount of text an AI model can see at one time.
That covers your instructions, the conversation so far, any files or tool results, and its own reply.
It’s measured in tokens, which are roughly three-quarters of a word each in English.
Whatever falls outside the window, the model can’t use.
Prompt vs context →
A prompt is the request you type. Context is everything the AI can see when it answers, including that request.
That means your files, examples, saved instructions, earlier messages and anything its tools fetched.
When an answer comes back wrong, the cause is usually missing or wrong context, not clumsy wording.
Claude and its tools
Claude Cowork →
Claude Cowork is the way of using Claude where you hand over a whole task instead of asking a question.
You point it at a folder and any apps you’ve connected, describe the result, and it plans the steps, works through them and saves a finished file.
Anthropic has since folded Cowork into Claude itself, so you may not see a separate tab.
Claude skills →
A Claude skill is a folder of instructions, sometimes with scripts and reference files, that Claude loads when your request matches it.
You write a process down once.
Claude follows it every time that job comes up, without you pasting the steps again.
MCP (Model Context Protocol) →
MCP, the Model Context Protocol, is an open standard that lets AI apps like Claude connect to other software: your email, calendar, files or database.
Anthropic released it in November 2024 and other AI apps have adopted it since.
When you add a connector in Claude and log in, you’re usually using MCP.
Claude skills vs MCP →
MCP connects Claude to your tools and data, so it can read your calendar or post in Slack.
A skill teaches Claude how you do a job, step by step. One gives it hands. The other gives it your habits.
Most useful setups need both.
Content and social media
Content pillar →
A content pillar is a recurring theme your posts keep returning to.
You choose it because your audience cares about it and it leads somewhere useful for your business.
Each post belongs to one pillar. Most accounts run on three to five.
A pillar earns its place with results, not because it’s easy to make.
Social media hook →
A social media hook is the first line or first frame of a post, the part that decides whether someone stops scrolling or keeps going.
On a slideshow it’s the text on slide one. In a video it’s the first sentence.
A good hook carries the whole thought, so the rest of the post only has to deliver on it.
Posting cadence →
Posting cadence is how often, and how regularly, an account publishes: one post a day, three a week, two a day at set times.
It’s a rhythm your audience and the platform learn to expect.
The right cadence is the one you can keep for months, not the one you can hit in a good week.
Faceless content →
Faceless content is social media content that never shows the creator’s face.
That means text over photos, slideshows, screen recordings, hands or a voiceover.
It lets you post often without filming yourself.
It reaches fast and builds trust slowly, so I treat it as one lane, not the whole road.
Content repurposing →
Content repurposing is turning one piece of content into several new ones, each shaped for a different place.
For example, a recorded talk becomes a week of posts.
Done well, every new piece stands on its own. Done badly, it’s the same summary pasted five times.
Client work
Creative brief →
A creative brief is a short document that tells a designer, writer or AI tool what to make, for whom and why, before anyone starts making it.
It names the goal, the audience, the message, the limits and what good looks like.
The brief decides what everyone is solving, so a weak one sets the whole project up to miss.
Coaching session notes →
Coaching session notes are the record a coach keeps of each session.
What the client brought, what shifted, what they committed to and where to pick up next time.
Good notes are short and keep the client’s own words. They exist for continuity, not as a transcript of everything said.
Ready to try something?
Choose one repeatable task. Add the context. Build a useful first version.