General AI assistants
Does: answers questions, writes, summarises, translates, brainstorms, analyses files.
Examples: Claude, ChatGPT, Gemini, Microsoft Copilot.
Best for: daily work, learning, drafting.
01 basics
What are AI tools?
An AI tool is software that uses an artificial intelligence model to perform tasks that normally need human intelligence: understanding language, recognising images, writing, translating, coding, or making predictions. Instead of following only fixed rules, the tool learned patterns from huge amounts of data and applies them to your input.
Simple view: an AI tool = an interface (chat box, app, plugin) + an AI model (the “brain”) + sometimes extra features like search, file upload or integrations.
Traditional software
You write exact rules. The same input always gives the same output. It breaks when something unexpected appears.
AI powered software
It learns patterns from data. It can handle new, messy input such as free text or photos, but results can vary and may sometimes be wrong.
Key terms
- Model
- The trained system that produces the output. Tools are built on top of models.
- LLM
- Large Language Model. Trained on massive text to understand and generate language. Powers most chat assistants.
- Token
- A small piece of text (a word or part of a word). Models read and write in tokens, and usage is often measured in them.
- Context window
- How much text the model can “keep in mind” in one conversation. Bigger windows handle longer documents.
- Training vs. inference
- Training is when the model learns from data. Inference is when it answers your prompt.
- Fine tuning
- Extra training on specific data so the model becomes better at a narrow job.
- RAG
- Retrieval Augmented Generation. The tool first looks up relevant documents, then the model answers using them, which improves accuracy.
- Multimodal
- A model that works with more than one type of data: text, images, audio or video.
- Hallucination
- When the AI states something false in a confident way. Always verify important facts.
- AI agent
- A system that plans steps and uses tools (search, code, apps) to reach a goal with less step by step guidance.
02 under the hood
How an AI tool works
Here is what happens, in simplified form, when you send a prompt to a text based AI tool.
- InputYou type a prompt or upload a file.
- TokeniseYour text is split into tokens and turned into numbers.
- ModelA neural network (usually a transformer) weighs how all tokens relate to each other.
- PredictIt predicts the most likely next token, again and again, building the reply.
- OutputTokens are turned back into readable text, an image or audio.
Why prompts matter
The model only sees what you give it. Clear instructions, context and examples lead to far better results than vague requests.
Why results can differ
Generation involves controlled randomness, so the same prompt can return different answers. This helps creativity but hurts exact repeatability.
Where the knowledge comes from
From training data plus anything you provide or the tool retrieves (web search, your files). Without search, a model may not know recent events.
Image and video tools
Most use diffusion models: they start from noise and gradually shape it into an image that matches your text description.
03 categories
Types of AI tools
Tools below are examples, not a ranking. Features change quickly, so always check each official website.
Search and research AI
Does: finds sources and gives answers with citations, or analyses your own documents.
Examples: Perplexity, NotebookLM, Elicit.
Best for: studying, reports, literature review.
Writing and editing
Does: grammar, tone, rewriting, marketing copy, email help.
Examples: Grammarly, Jasper, Notion AI.
Best for: content teams, students, professionals.
Image generation
Does: creates or edits pictures from text prompts.
Examples: Midjourney, DALL·E, Stable Diffusion, Adobe Firefly, Ideogram.
Best for: concept art, social posts, mockups.
Video generation and editing
Does: makes short clips from text or images, removes backgrounds, auto captions.
Examples: Runway, Pika, Descript, CapCut AI.
Best for: creators and marketers.
Voice, music and transcription
Does: text to speech, voice cloning, music creation, speech to text.
Examples: ElevenLabs, Suno, Whisper, Otter.
Best for: podcasts, accessibility, meetings.
Coding assistants
Does: suggests code, explains errors, writes tests, edits whole projects.
Examples: GitHub Copilot, Cursor, Claude Code.
Best for: developers and learners.
Design and presentation
Does: builds slides, layouts, logos and graphics from a short brief.
Examples: Canva, Figma AI features, Gamma.
Best for: quick visuals without a designer.
Meetings and notes
Does: records, transcribes and summarises meetings, extracts action items.
Examples: Otter, Fireflies, Notion AI.
Best for: teams with many meetings.
Workflow and agents
Does: connects apps and adds AI steps so tasks run by themselves.
Examples: Zapier, Make, n8n.
Best for: repetitive business tasks.
04 quick compare
Popular tools at a glance
| Tool | Type | Known for | Keep in mind |
|---|---|---|---|
| Claude | Chat assistant | Long documents, careful writing, coding help | Verify facts on important work |
| ChatGPT | Chat assistant | Versatile, large ecosystem of plugins and apps | Can state wrong answers confidently |
| Gemini | Chat assistant | Google Workspace integration, multimodal | Best value if you use Google apps |
| Perplexity | AI search | Answers with source links | Check the sources it cites |
| Midjourney | Image | Artistic, stylised images | Check licence and usage terms |
| Stable Diffusion | Image | Open source, runs locally, very customisable | More technical setup |
| ElevenLabs | Voice | Realistic text to speech | Never clone a voice without consent |
| GitHub Copilot | Coding | Code suggestions inside your editor | Review every line it writes |
| Zapier / Make / n8n | Automation | Connect apps with AI steps | Monitor workflows after launch |
05 decision guide
How to choose the right tool
Ask these questions
- What exact task do I need done?
- What type of output: text, image, audio, code?
- Does it connect to apps I already use?
- How accurate must the result be?
- What does it cost at my expected usage?
- How does it handle my data and privacy?
Smart starting strategy
- Start with one general assistant and learn it well
- Add a specialised tool only when you hit a clear limit
- Try the free plan before paying
- Test with the same prompt on two tools and compare
- Keep a note of prompts that work
06 prompting
Get better results: the RTCF prompt
Give the AI four things: Role, Task, Context and Format.
Role: You are a friendly career coach.
Task: Create a 4 week plan to learn web development.
Context: I am a complete beginner, 1 hour per day.
Format: A table with Week, Topics, and Practice project.Tip: if the answer is not good, do not start over. Reply with what to change, for example “Make it shorter and use simpler words.”
07 responsible use
Limits, risks and safe use
Accuracy
AI can hallucinate. Check names, numbers, dates and legal, medical or financial claims against reliable sources.
Privacy
Do not paste passwords, ID numbers or confidential client data. Read each tool’s data and training policy.
Copyright and ownership
Rules for AI generated content vary by country and tool. Check licence terms before commercial use.
Bias
Models learn from human data and can reflect its biases. Review output that affects people.
Over reliance
Use AI as an assistant, not a replacement for thinking. Keep learning the basics yourself.
Honesty
Be transparent when AI helped with school, work or published content, and follow the rules of your organisation.
08 roadmap
Your learning path
- Week 1Use one chat assistant daily. Learn prompting.
- Week 2Try tools in 3 categories: research, image, writing.
- Week 3Build a small workflow with Zapier or Make.
- Week 4Pick a project: a portfolio page, a content plan, or a study system.
- LaterLearn APIs and basic Python or JavaScript to build your own tools.
09 faq
Frequently asked questions
Are AI tools free?
Most have a free tier with limits. Heavier use, newer models and extra features usually need a paid plan.
Which AI tool is the best?
There is no single best. It depends on the task. Compare two or three on your own real examples.
Will AI replace my job?
AI automates tasks more than whole jobs. People who learn to work with AI tools usually gain an advantage.
Do I need to learn coding?
Not to use most AI tools. Coding helps if you want to build custom tools or automate deeply.
Is my data safe?
It depends on the provider and plan. Read the privacy policy and avoid sharing sensitive information.