Claude
AI
Training.

Chat is the core of Claude — a conversational interface where you type a question, request, or task and Claude responds instantly. Think of it as a highly capable colleague available 24/7 who never gets tired, never judges your question, and always has time for you.

You talk to Claude the way you would talk to a person. Ask questions, give instructions, think out loud, or say "help me write this." Claude reads what you type, understands the context, and responds. You can then reply, correct it, or go deeper. It remembers everything said within that session.

Projects are dedicated workspaces inside Claude where you can organise all your conversations, files, and instructions around a specific piece of work. Instead of starting fresh every time, Claude remembers the context of everything in that project — permanently.

Inside a Project, you can upload documents, set standing instructions, and have multiple conversations — all connected. Claude always knows the background. You do not have to re-explain who you are, what you are working on, or what the rules are every time you open a new chat.

A prompt is the instruction you give Claude. The quality of what Claude produces depends almost entirely on the quality of what you ask. Prompting is the skill of writing clear, structured instructions that get you the output you actually need — first time.

Good prompting removes ambiguity. Instead of Claude guessing what you want, a well-structured prompt tells it your role, the task, the context, the audience, and the format — so the response is precise, relevant, and usable.

Artifacts are live, visual outputs that Claude generates and displays directly alongside your conversation. Instead of just giving you text, Claude builds something you can see, interact with, and use immediately — a table, a chart, a form, a dashboard, a document, or a webpage.

When you ask Claude to create something visual or structured, it renders it in a dedicated panel next to the chat. You can view it, copy it, download it, or ask Claude to revise it — all without leaving the conversation. No design tools, no coding, no formatting from scratch.

Cowork is Claude's desktop application — a version of Claude that lives on your computer and works directly with your files, folders, and local systems. It is the most powerful way to use Claude for day-to-day professional work because it connects Claude to your actual working environment.

In Cowork, Claude can read, write, and edit files on your computer. You connect a folder — your workspace — and Claude works inside it. It can open documents, save outputs, maintain a memory of who you are, and run automated tasks. Everything you build in Cowork persists across every session.

Skills are installable capability packs that extend what Claude can do inside Cowork. With the right Skills installed, Claude becomes a specialist — one that can build Excel models, create Word documents, design presentations, generate PDFs, and more, all to professional standard.

Each Skill gives Claude deep, domain-specific capability for a particular type of output. When you ask Claude to produce something that requires a Skill, it follows a structured, expert-level process — producing a properly formatted, ready-to-use file, not just text on a screen.

Live Artifacts are persistent, interactive pages that Claude builds and saves — pages you can open again and again, and that pull fresh data every time you do. Unlike a regular Artifact, a Live Artifact is a living page that stays current with your connected tools and data sources.

Claude builds the page once. Every time you open it, it fetches the latest data from your connected systems — email, calendar, SharePoint, project tools — and displays an up-to-date view. You do not need to ask Claude again. The page does it for you.

Scheduled Tasks let you set Claude to run a task automatically — at a time you choose, on a repeat you define. You set it up once, and Claude does the work at the right time without you having to remember, open Claude, or type anything.

You tell Claude what to do and when to do it. Claude runs the task at the scheduled time — whether you are at your desk or not — and delivers the output. It can pull data from connected tools, generate reports, draft summaries, or complete any repeating task you currently do manually.

Dispatch is Claude's ability to break a large, complex task into multiple workstreams and run them simultaneously — each handled by a dedicated sub-agent working in parallel. Instead of completing steps one after another, Claude orchestrates several agents at once, like a manager delegating to a team.

When you give Claude a multi-part task, Dispatch splits the work and runs each part concurrently. The results come back as a single unified output. What might take Claude several minutes done sequentially can be completed in a fraction of the time.

Connectors are integrations that link Claude directly to the tools and platforms Raffles already uses — Microsoft 365, Google Workspace, SharePoint, Outlook, Teams, and more. Once connected, Claude can read from and work with your live systems without you copying, pasting, or uploading anything.

With a Connector active, Claude reaches into your connected platform and pulls the information it needs in real time. It can read your emails, check your calendar, search SharePoint, pull documents from OneDrive, and more — then act on what it finds.

Plugins are installable bundles that extend Claude with a complete set of capabilities in one go. Where a Skill adds one specific capability, a Plugin adds an entire ecosystem — multiple skills, connectors, tools, and instructions — packaged together for a specific role or organisation.

You install a Plugin and Claude gains everything inside it at once. Plugins built for Raffles carry Raffles-specific knowledge, brand rules, and workflows baked in from the start. Every staff member who installs the Raffles Plugin starts with a Claude that already knows the organisation.

Memory is Claude's built-in ability to remember facts, preferences, and context across conversations — at the account level. It is personal to you, always active, and grows over time as you use Claude.

When something important comes up in a conversation — a preference, a decision, a key fact — Claude can save it to Memory. The next time you open Claude, even in a completely new chat, it already knows. Memory persists indefinitely until you edit or delete it.

Claude can read, analyse, and act on files you upload — PDFs, Word documents, Excel spreadsheets, CSVs, images, and more. You upload the file and tell Claude what to do with it. Claude reads the full content and responds based on what is actually in the document.

Claude can summarise a document, extract specific data, compare two files, answer questions about the content, reformat information, spot inconsistencies, and produce new outputs based on what it reads. It handles documents the way a skilled analyst would — quickly, thoroughly, and without missing detail.

Vision is Claude's ability to analyse images and visual content. You share an image — a screenshot, a photograph, a scanned document, a chart, a diagram — and Claude describes, interprets, extracts, or acts on what it sees.

Claude reads visual content the way it reads text. It can describe what is in an image, extract text from a scanned document, interpret a chart or graph, identify issues in a screenshot, review a design, and answer questions about what it sees.

Web Search gives Claude the ability to search the internet in real time during your conversation. Instead of relying only on what you tell it, Claude goes out, finds current information, and uses it to answer your question or complete your task.

When you ask Claude something that requires up-to-date information — current regulations, recent news, live pricing, a competitor's latest announcement — Claude runs a web search, reads the relevant results, and incorporates what it finds into its response. Every answer comes with cited sources.

Deep Research is Claude's advanced research capability — a mode where it conducts structured, multi-source investigation across the web, critically evaluates what it finds, and produces a comprehensive, cited report. It is a full research process, compressed into minutes.

Claude fans out across multiple sources simultaneously, cross-checks claims, discards unreliable information, and synthesises everything into a single, well-structured report with citations. It approaches the task the way a trained researcher would — looking for consensus and flagging contradictions.

Claude in Chrome is a browser-based agent that operates directly inside Google Chrome. It can navigate websites, read page content, fill information, interact with web-based systems, and complete multi-step browser tasks on your behalf.

You describe what you want done on the web and Claude does it. It opens pages, reads content, extracts data, moves through workflows, and reports back with what it found or what it completed. It sees the web the way you do — rendered pages, not just raw code.