These AI tool tutorials show you how to use the assistants people actually open at work, with a ten-minute starter exercise, techniques that suit each tool and a plain account of what the vendor says about your data. They are free to read and written independently: Tosheo.ai is not affiliated with OpenAI, Anthropic, Google, Microsoft, Perplexity AI or Canva.
ChatGPT is OpenAI's general-purpose AI assistant: you type or speak a request, attach files if you need to, and it writes, summarises, analyses or researches in reply. This ChatGPT tutorial for beginners teaches you how to give it a work brief, check what it returns and keep your work organised with Projects.
Open the tutorialClaude is Anthropic's AI assistant for writing, analysis and research, available on the web, on desktop and on mobile. This Claude tutorial teaches you to work through a long document with it, keep reference material in a Project and turn the result into an artifact you can reuse, which is the quickest way to learn Claude for real work.
Open the tutorialGemini is Google's AI assistant, used at gemini.google.com, in the mobile app and, on qualifying Workspace plans, with a work account. This Gemini tutorial shows you how to use Google Gemini for research and drafting, how to save a reusable Gem and how to steer Deep Research before it runs.
Open the tutorialMicrosoft Copilot is Microsoft's AI assistant, and it behaves differently depending on whether you sign in with a personal Microsoft account or a work or school account. This Microsoft Copilot tutorial explains both, then teaches you to ground Copilot Chat in your own work files and check what it produces.
Open the tutorialCopilot in Excel is Microsoft's AI assistant inside the Excel workbook: you describe what you want in plain language and it edits sheets, writes formulas, builds charts and PivotTables or answers questions about your data. This tutorial teaches you to choose the right mode, review each change and check the numbers before anyone relies on them.
Open the tutorialNotebookLM is Google's AI research notebook that answers questions only from sources you add, with inline citations; Google renamed it Gemini Notebook in July 2026. This NotebookLM tutorial teaches you to build a notebook from documents you are allowed to use, question them, and turn them into briefings and Audio Overviews you can check line by line.
Open the tutorialPerplexity is an AI search tool from Perplexity AI that searches the web in real time and answers with numbered citations you can open. This tutorial teaches you to ask research questions it can answer well, follow up within a session and check every source before a finding goes into your work.
Open the tutorialCanva AI is the set of AI features Canva groups as Magic Studio, which generate designs, write and rewrite copy, edit images and reformat a design for other channels inside the Canva editor. This tutorial teaches you to produce an on-brief design quickly, then check the copy, the image rights and the facts before it goes out.
Open the tutorialGrammarly is an AI writing assistant that checks and rewrites your text inside the apps where you already write, and its official site says it is now part of Superhuman. This tutorial shows how to set it up for work, edit a real message in ten minutes, steer its generative AI rewrites and check what its data settings mean before you rely on it.
Open the tutorialMidjourney is an AI image and short video generator that turns a written prompt into images you can vary, edit and upscale on its website or in Discord. This tutorial covers how to write a prompt it understands, which parameters matter for work, a ten-minute poster example, and what its terms say about privacy, ownership and real people.
Open the tutorialOtter.ai is an AI meeting assistant that records and transcribes meetings, identifies speakers and writes a summary with decisions and action items. This tutorial shows how to set it up for Zoom, Teams or Google Meet, run a ten-minute check on its output, and handle consent and data properly.
Open the tutorialAdobe Firefly is Adobe's generative AI app for creating and editing images, video and audio from prompts, with Adobe's own models and a choice of partner models in one place. This tutorial shows how to generate and edit a work image with composition and style references, and explains what Adobe says about commercial use, training data and Content Credentials.
Open the tutorialNotion AI is the AI built into Notion that writes on your pages, answers questions from your workspace and connected apps, and runs multi-step tasks through Notion Agent. This tutorial shows how to turn scattered notes into a project brief and task database, how to get grounded answers rather than general ones, and what Notion says about permissions and data.
Open the tutorialElevenLabs is an AI audio platform that turns text into natural-sounding speech and also offers voice cloning, dubbing, transcription, sound effects and music. This tutorial shows how to produce a short training voiceover, control pauses and pronunciation, and stay within ElevenLabs' rules on consent, voice cloning and commercial use.
Open the tutorialGamma is an AI tool for creating presentations, documents, webpages and social content from a prompt, your own notes or an imported file. This tutorial shows how to build a short client briefing from your own text, edit it with AI without losing your words, export it to PowerPoint and set Gamma's data controls.
Open the tutorialOpenAI dots are always-on agents inside ChatGPT that keep working towards a goal between conversations, using their own cloud computer and the apps you connect. This tutorial shows you how to set up a ChatGPT dot, give it a first project with clear approval rules and review its work.
Open the tutorialGrok Bot is SpaceXAI's beta product for persistent AI teammates, called Bots, that work on a shared cloud computer and finish tasks in your real apps. This tutorial shows you how to create a focused Bot, hand it a first task with clear limits and turn a reliable process into a skill.
Open the tutorialMeta Muse is Meta's personal AI agent, an app that takes actions for you on its own secure cloud computer rather than only answering questions. This tutorial shows you how to start a Muse conversation, give the Meta Muse agent a bounded task, set what it may do and check its audit trail before you rely on it.
Open the tutorialMicrosoft Copilot Studio is a low-code studio for building AI agents and workflows that connect to your organisation's data and run in Teams, Microsoft Copilot, websites and other channels. This Copilot Studio tutorial walks you through describing an agent, grounding it in knowledge, testing it properly and understanding what your licence lets you publish.
Open the tutorialA custom GPT is a version of ChatGPT set up for one purpose with its own instructions, knowledge files and tools. OpenAI is retiring custom GPTs in favour of plugins, so this tutorial shows how to create a custom GPT where your workspace still allows it, how to design one that works and how to migrate it to a plugin.
Open the tutorialGemini Deep Research is a feature in Google's Gemini app that turns your question into a research plan, browses many sources and returns a multi-page report in a few minutes. This tutorial shows you how to brief a run, edit the plan before it starts, check the sources behind the report and export it, using a fictional market-entry question.
Open the tutorialZapier automation connects the apps you already use so that an event in one, such as a new form entry, makes things happen in others without anyone copying data across. This tutorial builds one fictional workflow from trigger to tested, multi-step Zap with an AI step, and shows how to handle errors before they reach a customer.
Open the tutorialZapier Agents are AI assistants that carry out work across your connected apps, deciding which tools to use to reach a goal you describe, rather than following fixed steps. Zapier is moving this capability into its AI by Zapier step inside the Zap editor, which adds per-tool approvals, so this tutorial covers both and shows how to review what an agent did.
Open the tutorialYour diagnostic picks the order and skips what you already know. Start any lesson with a 7-day free trial, no credit card needed.
Every tutorial follows the same order, so you can compare tools without hunting for the part you need. It starts with what the tool is good for at work, then how to get an account and find it, then a starter exercise built around a fictional business so you never need to paste real company material to practise.
After the starter, each page gives techniques specific to that tool, such as Projects in ChatGPT and Claude, Gems in Gemini, Plan mode in Copilot in Excel or source citations in Perplexity. Then it covers what not to paste in, what the vendor's own pages say about data controls, and where the tool falls short. Every capability is taken from the vendor's official product or help pages, which are listed with the date we checked them.
A tutorial teaches the tool; a Tosheo.ai path teaches you to use it for your own job. After a short diagnostic, each learner gets an adaptive path of 10 to 15 minute lessons matched to their role and level. Each lesson has a short explanation, a worked example from that role and a hands-on task in a real AI tool, marked against a visible rubric with specific feedback.
Paths start with AI at work and continue into data science, machine learning, AI and blockchain. Every subject starts with concepts and no-code practice, and an optional Code track adds hands-on Python, SQL, AI APIs and agents. It is not a full-time bootcamp or a degree. Every plan starts with a 7-day free trial, and no credit card is needed.
Start with the tool your employer already approves. If your company runs Microsoft 365, read the Microsoft Copilot tutorial first; if it runs Google Workspace, start with Gemini. If you are choosing for yourself, the ChatGPT and Claude tutorials are the broadest starting points, and the comparison in our AI tools at work guide helps you decide.
Yes. Every tutorial in the library is free to read with no account. They are written by Tosheo.ai, which is independent of the tool vendors and is not paid by them.
Each capability is taken from the vendor's official product or help pages, and each page lists those sources with the date they were checked. Menus and plan features change often, so if a step does not match your screen, the vendor's help page is the final word.
No. Every starter exercise uses a fictional company and invented details, so you can practise without sharing anything real. When you move to real work, follow your employer's AI policy on what may go into each tool.
Tosheo.ai turns the same tools into a personalised path for your role, with hands-on tasks marked against a visible rubric and specific feedback. Team managers see skill progress and a team report, never raw chat logs.
Explore role-specific lessons, practical tasks and the manager view.
