Every workflow in this section is the same move: start with an ask, then stack on context, search, images, and video. More ingredients in, better answers out.
Ask Something!
"How do I bake a cake?"
Your question goes in as plain language. The model predicts the most likely helpful answer, one word at a time.
No magic words, no special syntax. You ask in plain language and the model writes back the most likely helpful answer — ingredients, equipment, steps. The catch: with nothing else to go on, it has to answer for the average person asking that question.
Ask Something + Context!
"How do I bake a cake? I'm lactose intolerant and I love chocolate!"
Two facts about you, and the same question produces a different cake: dairy-free and chocolate. Context is the cheapest upgrade there is.
Same question, two extra facts about you — and the whole answer changes: milk and butter get swapped for plant-based alternatives, and the recipe turns chocolate. Context is the cheapest upgrade in AI. Every detail you share reshapes the entire answer.
Ask Something + Search!
"How do I bake a cake? I'm lactose intolerant and I love chocolate! Give me the best rated recipes!"
Models stop learning at their training cutoff. Search lets them read today's web — real recipes, real ratings, real sources it can cite.
A model's built-in knowledge stops at its training cutoff. Asking for "the best rated" flips on a different move: the model searches the live web, reads real recipe pages and reviews, and answers with sources it can cite. Most major chatbots have a search or "browse" toggle — use it whenever the answer depends on what's true today.
Ask Something + Image Context!
What can multimodal models do?
Frontier models don't just read — they look. Text, images, and files can all go into a single prompt.
Most modern frontier models are what we call multimodal models. That means you can use as input not just text, but also images, files, and even videos with some of them. Model names change every few months, so instead of memorizing them, check our model table to see which model is currently best for each use case.
Ask Something + Video Context!
How can I analyze videos with AI?
Drop in a video and the model scans frames and audio — transcripts, summaries, and timestamped answers come out.
Some frontier models are especially good at analyzing videos, giving you transcripts, summaries and more! Look for the models with video input support in our model table.
Visualize Data!
How can I visualize data with AI?
Paste raw rows, get a chart in minutes instead of an afternoon in a spreadsheet — then export it right back to CSV.
Most major chatbots (you'll find the current best ones in our model table) have a tool called Canvas that can help you visualize almost every kind of data in a matter of minutes instead of losing hours in an excel sheet. And you can ask it to turn those visualizations into excel data, csv or whatever format you need!
Create Images!
"A cool image of a flying bottle!"
Image models start from pure noise and sculpt it toward your words until a picture appears.
AI is incredible at making images, and new and better models are coming week after week — so any name we print here would be outdated in a month. Check the image generation section of our model table for the current best picks. The top models give you professional output with well-written, correctly spelled text inside the image.
Animate an Image!
How can I animate images with AI?
Video models take one still and imagine every frame around it. Press play.
There are tons of State of the Art (SOTA) video generation models at the moment, and the ranking changes monthly — see the video generation section of our model table for the current favourites.
How to make a Presentation
"Create a presentation for a tech company that sells AI sparkling water bottles!"
One sentence becomes a structured deck: problem, product, market, business model, the ask.
One sentence in, a structured deck out: problem, product, market, business model, and the ask — with consistent design and one point per slide. Check out this example of an AI agent that can help you create presentations: GenSpark AI Presentation Assistant
Ask Something with Context! on Steroids (Projects)
What are AI Projects and how can I use them?
Upload your files and instructions once. Every new chat in the project starts with all of it already in mind.
Now let's talk about one of the most powerful features any worker can exploit in almost any AI chatbot: Projects. Projects are the culmination of everything we explained, because they are a mix of:
- Good Context
- Good Prompting
Almost every chatbot has a functionality called Projects (or similar):
- ChatGPT (Projects)
- Claude (Projects)
- Grok (Projects)
- Gemini (Gems)
- NotebookLM (Notebooks)
- Perplexity (Spaces)
There you can put all your relevant files: PDFs, Markdown files (best choice for AI) — and in some, like NotebookLM, you can add websites, Google Docs, YouTube videos and more!