How FormelyAI Works

1. Describe your goal

Start by telling the generator what you want to achieve in plain language — a formula, an email, a report outline, a slide deck or anything else. You then choose your target AI model (ChatGPT, Gemini, Claude, Copilot, DeepSeek, Grok, Perplexity, Mistral or Qwen), the task type, the desired tone, the audience's experience level, the output format and the language.

You do not need to know prompt-engineering jargon. The selectors translate those choices into the structured instructions the model needs, so a beginner can produce a professional-grade brief on the first try.

2. Generate an optimized prompt

Our system applies proven prompt-engineering frameworks — role, context, constraints and format — to assemble a complete prompt tailored to your chosen model and settings. The result is a single, copy-ready instruction that removes the guesswork the model would otherwise have to fill in.

Because each model interprets phrasing slightly differently, the generator tunes the wording to the platform you selected, giving you a stronger starting point than a generic one-size-fits-all prompt.

3. Copy, run and refine

Paste the generated prompt into ChatGPT, Gemini, Claude, Copilot or any other model and run it. Treat the first response as a strong draft, then refine iteratively — ask for a shorter version, a different tone or an added section — using the tips provided alongside each tool.

Iterating on a good prompt is far faster than restarting from a vague one. This loop is where most of the time savings come from, and it is a habit we encourage throughout the Learning Center.

Built to scale

New generators, prompt categories and articles are added regularly at no extra cost. The platform is designed so new tools slot in without any disruption to your workflow, and the Prompt Library and Office AI hub grow alongside the Learning Center so there is always something new to apply.