> ## Documentation Index > Fetch the complete documentation index at: https://notes.kodekloud.com/llms.txt > Use this file to discover all available pages before exploring further. # Your first prompt > Guide showing how to craft an effective first prompt by specifying persona, context, and task to get fast useful AI responses In this lesson you'll learn how to write your first effective prompt. Twelve minutes before Maya’s next client call, she’s been avoiding a follow-up email for two weeks. The lead—a bakery owner she met about Instagram captions—hasn’t replied. Each day she waits, the opportunity drifts further away. DeepSeek is open in the next tab. This is the moment of truth: will it save her time, or will it just waste more of it? The chat interface is intentionally minimal. At the top left, a "New chat" button; a column for past conversations (empty here); a model selector at the top; and a large input area at the bottom where you type. A minimalist chat UI mockup with the heading "No tabs. No settings. Just four things." It shows a sidebar with "New chat" and "Past chats", a main panel labeled "V4‑Flash", and a "Type here..." input field. No tabs, no hidden settings to fiddle with — just a focused chat prompt box. Maya types her first message into the input field. Writing it as one short, complete instruction helps the model deliver an immediate, usable result. ```text theme={null} "I'm a freelance content writer. Two weeks ago, I had a meeting with a bakery owner about helping with her Instagram captions. She hasn't replied. Write a short follow-up email that sounds warm and patient, not pushy. Sign it, Maya." ``` She presses Enter. DeepSeek pauses briefly, then streams the reply in real time. Before she finishes reading the first sentence, the model is on the third. Forty words later she has a three-line email she can paste straight into Gmail. The tone is right, and her name is at the bottom. She changes just one phrase and hits send. What had stalled in her inbox for two weeks took ninety seconds with DeepSeek. A screenshot of a webpage titled "From blank page to a draft she can paste." showing a short follow-up email about Instagram captions for a bakery signed "Maya." Below the email are three stat boxes (40 words, 1 phrase changed, 90s blank to sent) and the line "A two-week stall, cleared in ninety seconds." Look at why that single message worked. Maya explicitly told DeepSeek: * Who she is (persona), * The relevant context (a real meeting, a real prospect, no reply), and * The exact task and tone (short follow-up email, warm and patient), plus how to sign it. That combination gave the model enough information to produce a clean, nearly final draft on the first try. An infographic titled "Three invisible moves — Why that one message worked" showing three numbered cards: 1) "Who she is" with the line "I'm a freelance content writer.", 2) "The context" with "A real meeting, a real prospect, no reply.", and 3) "A specific job" with "Write a short follow-up email." A concise, effective prompt usually contains three parts: (1) the persona or role to adopt, (2) the relevant context or constraints, and (3) the specific task plus desired tone or format. The more of these you include up front, the less editing the result will need. Summary — The three essential prompt components: | Component | What to include | Example | | - | - | - | | Persona | Who or what role the model should adopt | `I'm a freelance content writer.` | | Context | Brief background, constraints, or relevant facts | `Two weeks ago I met a bakery owner about Instagram captions; she hasn't replied.` | | Task & Tone | The exact output you want and the tone/format | `Write a short follow-up email, warm and patient. Sign it Maya.` | Practical tips to get a strong first draft * Be explicit about the persona if voice matters (e.g., "as a junior developer" or "as a friendly small-business owner"). * Include only the context that affects the result—avoid long unrelated backstory. * Specify format and length (email, subject line, 2-3 bullets, 40–60 words). * Mention tone: warm, assertive, technical, playful, concise, etc. * If you want signatures, placeholders, or variables, state them in the prompt. A quick checklist before you hit Enter * [ ] Persona or role is specified (if needed) * [ ] Relevant context is clear and concise * [ ] Task, format, and tone are defined * [ ] Any output constraints (word count, sign-off, bullets) are included Maya’s prompt succeeded because it had shape: it told the model who, what, and how. Some prompts land immediately; others need iteration. Use the three invisible moves — persona, context, task — and you’ll get closer to a ready-to-send result on the first try. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.