[{"data":1,"prerenderedAt":462},["ShallowReactive",2],{"guide-en-prompting":3,"guide-siblings-en-prompting":285,"guide-alt-en-prompting":452},{"id":4,"title":5,"author":6,"body":7,"date":259,"description":260,"extension":261,"faq":262,"meta":275,"navigation":276,"order":277,"path":278,"readTime":279,"seo":280,"stem":281,"topic":282,"translationId":283,"updated":259,"__hash__":284},"guides\u002Fguides\u002Fprompting\u002Findex.md","What is a prompt? How to write prompts that work at work","Walma Engineering",{"type":8,"value":9,"toc":248},"minimark",[10,14,19,22,25,43,46,50,53,59,65,71,77,83,86,92,96,102,108,114,120,126,132,136,168,172,175,178,209,212,216,219,240],[11,12,13],"p",{},"A prompt is what you type to an AI model. That sounds too simple to need a guide, and yet the difference between a useless answer and a useful one is almost always in the prompt, not the model. This guide covers what a prompt is, what a good one contains, the techniques that consistently improve results, and how teams stop rewriting the same prompt every day.",[15,16,18],"h2",{"id":17},"what-a-prompt-is","What a prompt is",[11,20,21],{},"A prompt is the input the model turns into output. In a chat tool it is your message. In an application it is the message plus whatever the application adds around it: instructions, documents, conversation history, tool results. The model sees all of it as one long text and predicts what should come next.",[11,23,24],{},"Two layers matter in practice:",[26,27,28,37],"ul",{},[29,30,31,32,36],"li",{},"The ",[33,34,35],"strong",{},"system prompt"," sets standing instructions: who the model is, what it may and may not do, what tone and format to use. In company deployments this is usually written once, centrally.",[29,38,31,39,42],{},[33,40,41],{},"user prompt"," is the specific request. This is the part most people mean when they say \"prompt\".",[11,44,45],{},"Everything below applies to both, but most of your daily leverage is in the user prompt.",[15,47,49],{"id":48},"the-anatomy-of-a-good-prompt","The anatomy of a good prompt",[11,51,52],{},"Good prompts share a structure, whether or not the writer thinks about it. Five parts, in rough order of importance:",[11,54,55,58],{},[33,56,57],{},"1. The task."," What you want, as a verb. \"Summarise\", \"draft\", \"compare\", \"extract\", \"rewrite\". Vague tasks (\"look at this\") get vague answers.",[11,60,61,64],{},[33,62,63],{},"2. The context."," What the model needs to know to do the task well: the audience, the situation, the constraints, the source material. Models do not know your company, your customer or last week's meeting unless you tell them.",[11,66,67,70],{},[33,68,69],{},"3. The format."," What the output should look like: a table, five bullets, a 120-word paragraph, JSON with these fields. If you do not specify, you get the model's default, which is rarely what you would have chosen.",[11,72,73,76],{},[33,74,75],{},"4. An example."," One example of the output you expect is worth a paragraph of description. Models are extremely good at matching a pattern.",[11,78,79,82],{},[33,80,81],{},"5. The constraints."," What to avoid, what to include, how long, which sources to use, what to do when unsure. \"If the document does not say, write 'not stated' rather than guessing.\"",[11,84,85],{},"A prompt that has all five:",[87,88,89],"blockquote",{},[11,90,91],{},"Summarise the attached customer call transcript for our account manager. Context: the customer is evaluating our platform against a competitor and raised concerns about data residency. Format: three sections, \"Concerns raised\", \"Commitments we made\", \"Next steps\", each as bullets, maximum 150 words total. Example of tone: (paste a previous summary). If a next step has no owner or date in the transcript, mark it \"unassigned\".",[15,93,95],{"id":94},"techniques-that-reliably-help","Techniques that reliably help",[11,97,98,101],{},[33,99,100],{},"Give the model a role."," \"You are a senior contract lawyer reviewing for a Swedish SME\" changes vocabulary, depth and what the model considers important. Keep it specific; \"you are an expert\" alone does little.",[11,103,104,107],{},[33,105,106],{},"Show, don't describe."," Few-shot prompting means including two or three examples of input and expected output. For classification, extraction and formatting tasks it beats any description.",[11,109,110,113],{},[33,111,112],{},"Let it think first."," For reasoning tasks, ask the model to work through the problem before answering, or use a model's built-in extended thinking. It reduces confident wrong answers.",[11,115,116,119],{},[33,117,118],{},"Separate data from instructions."," Put pasted documents inside clear delimiters (triple quotes, XML-style tags) and refer to them by name. It keeps the model from confusing content with commands, which also matters for security.",[11,121,122,125],{},[33,123,124],{},"Ask for structure when a program reads the output."," JSON with a schema, or a fixed table. Most models can be told to return only the structure and nothing else.",[11,127,128,131],{},[33,129,130],{},"Iterate in the conversation."," The second prompt (\"shorter, and lead with the risk\") is often more effective than trying to get the first one perfect.",[15,133,135],{"id":134},"mistakes-that-ruin-prompts","Mistakes that ruin prompts",[26,137,138,144,150,156,162],{},[29,139,140,143],{},[33,141,142],{},"Asking for several things at once."," Split them, or number them and ask for numbered answers.",[29,145,146,149],{},[33,147,148],{},"Leaving out the audience."," A summary for the CFO and one for the engineering team are different documents.",[29,151,152,155],{},[33,153,154],{},"Assuming shared context."," The model was not in the meeting.",[29,157,158,161],{},[33,159,160],{},"Trusting the first answer for facts."," Ask for sources, or give the model the source material and tell it to use only that.",[29,163,164,167],{},[33,165,166],{},"Pasting confidential material into a consumer tool."," The prompt is data. Where it goes is a data protection question. Use company-approved tools that process data in the right region.",[15,169,171],{"id":170},"from-prompts-to-skills","From prompts to skills",[11,173,174],{},"Once a prompt works, the problem changes: how does the rest of the team get it? Copying two pages into every session does not scale, and the wiki page nobody reads is where good prompts go to die.",[11,176,177],{},"The answer that has emerged is to package prompts as reusable assets:",[26,179,180,186,197],{},[29,181,182,185],{},[33,183,184],{},"Prompt libraries"," in the tool itself (Projects in Claude, custom GPTs, Copilot agents).",[29,187,188,191,192,196],{},[33,189,190],{},"Instruction files"," that agents load automatically, such as ",[193,194,195],"code",{},"CLAUDE.md"," in a code repository.",[29,198,199,202,203,208],{},[33,200,201],{},"Skills",": folders with instructions, examples and scripts that an agent loads when a task matches. This is the current best practice for coding agents and increasingly for knowledge work. Our ",[204,205,207],"a",{"href":206},"\u002Fen\u002Fguides\u002Fskills","guide to Claude skills"," explains how they work.",[11,210,211],{},"At that point \"prompt engineering\" has become \"context engineering\": deciding what the model should have in front of it for each kind of task, and maintaining that centrally.",[15,213,215],{"id":214},"prompts-in-a-company","Prompts in a company",[11,217,218],{},"Three things are worth deciding centrally rather than leaving to each user:",[220,221,222,228,234],"ol",{},[29,223,224,227],{},[33,225,226],{},"The system prompt"," for each approved tool: tone, language, what the tool may claim, what it must refuse.",[29,229,230,233],{},[33,231,232],{},"The approved prompt library",": the prompts and skills that encode how your company does recurring tasks.",[29,235,236,239],{},[33,237,238],{},"Where prompts are processed."," A prompt with customer data in it is a data transfer. Company tools should run models in your region, and someone should be able to see what was sent.",[11,241,242,243,247],{},"Walma AI Hub gives companies one place to run Claude, GPT and the other leading models in their own EU region, with skills and prompts distributed centrally and every call logged. If you are standardising how your organisation prompts, ",[204,244,246],{"href":245},"\u002Fen\u002Fai-hub","book a walkthrough",".",{"title":249,"searchDepth":250,"depth":251,"links":252},"",2,3,[253,254,255,256,257,258],{"id":17,"depth":250,"text":18},{"id":48,"depth":250,"text":49},{"id":94,"depth":250,"text":95},{"id":134,"depth":250,"text":135},{"id":170,"depth":250,"text":171},{"id":214,"depth":250,"text":215},"2026-09-11","A prompt is the instruction you give an AI model. This guide explains what goes into a good one, the techniques that reliably improve answers, the mistakes that ruin them, and how teams move from ad-hoc prompts to reusable skills.","md",[263,266,269,272],{"q":264,"a":265},"What is a prompt?","A prompt is the text you give an AI model to tell it what you want: a question, an instruction, some context, and often an example of the output you expect. Everything the model produces is shaped by it.",{"q":267,"a":268},"What is prompt engineering?","Prompt engineering is the practice of writing prompts deliberately: choosing the role, task, context, format and examples so the model gives reliably good answers. It is less about magic phrases and more about being clear about what you actually want.",{"q":270,"a":271},"What is the difference between a system prompt and a user prompt?","The system prompt sets standing instructions for the whole conversation: who the model is, what it may do, how it should answer. The user prompt is the individual request. In company tools the system prompt is usually set centrally and users only write the request.",{"q":273,"a":274},"Does prompt engineering still matter with newer models?","Less for tricks, more for clarity. Modern models need fewer workarounds, but they still cannot read your mind. Clear goals, relevant context and an example of the expected output matter as much as ever, and for agents the prompt has grown into skills and instruction files.",{},true,0,"\u002Fguides\u002Fprompting","9 min read",{"title":5,"description":260},"guides\u002Fprompting\u002Findex","prompting","prompt-guide","-2H4J7hhBWFd0LkaPMeJfka1ljMAeBdL5iucggvbjn4",[286],{"id":4,"title":5,"author":6,"body":287,"date":259,"description":260,"extension":261,"faq":445,"meta":450,"navigation":276,"order":277,"path":278,"readTime":279,"seo":451,"stem":281,"topic":282,"translationId":283,"updated":259,"__hash__":284},{"type":8,"value":288,"toc":437},[289,291,293,295,297,307,309,311,313,317,321,325,329,333,335,339,341,345,349,353,357,361,365,367,389,391,393,395,413,415,417,419,433],[11,290,13],{},[15,292,18],{"id":17},[11,294,21],{},[11,296,24],{},[26,298,299,303],{},[29,300,31,301,36],{},[33,302,35],{},[29,304,31,305,42],{},[33,306,41],{},[11,308,45],{},[15,310,49],{"id":48},[11,312,52],{},[11,314,315,58],{},[33,316,57],{},[11,318,319,64],{},[33,320,63],{},[11,322,323,70],{},[33,324,69],{},[11,326,327,76],{},[33,328,75],{},[11,330,331,82],{},[33,332,81],{},[11,334,85],{},[87,336,337],{},[11,338,91],{},[15,340,95],{"id":94},[11,342,343,101],{},[33,344,100],{},[11,346,347,107],{},[33,348,106],{},[11,350,351,113],{},[33,352,112],{},[11,354,355,119],{},[33,356,118],{},[11,358,359,125],{},[33,360,124],{},[11,362,363,131],{},[33,364,130],{},[15,366,135],{"id":134},[26,368,369,373,377,381,385],{},[29,370,371,143],{},[33,372,142],{},[29,374,375,149],{},[33,376,148],{},[29,378,379,155],{},[33,380,154],{},[29,382,383,161],{},[33,384,160],{},[29,386,387,167],{},[33,388,166],{},[15,390,171],{"id":170},[11,392,174],{},[11,394,177],{},[26,396,397,401,407],{},[29,398,399,185],{},[33,400,184],{},[29,402,403,191,405,196],{},[33,404,190],{},[193,406,195],{},[29,408,409,202,411,208],{},[33,410,201],{},[204,412,207],{"href":206},[11,414,211],{},[15,416,215],{"id":214},[11,418,218],{},[220,420,421,425,429],{},[29,422,423,227],{},[33,424,226],{},[29,426,427,233],{},[33,428,232],{},[29,430,431,239],{},[33,432,238],{},[11,434,242,435,247],{},[204,436,246],{"href":245},{"title":249,"searchDepth":250,"depth":251,"links":438},[439,440,441,442,443,444],{"id":17,"depth":250,"text":18},{"id":48,"depth":250,"text":49},{"id":94,"depth":250,"text":95},{"id":134,"depth":250,"text":135},{"id":170,"depth":250,"text":171},{"id":214,"depth":250,"text":215},[446,447,448,449],{"q":264,"a":265},{"q":267,"a":268},{"q":270,"a":271},{"q":273,"a":274},{},{"title":5,"description":260},[453,456,459],{"loc":454,"href":455},"sv","\u002Fguider\u002Fprompting",{"loc":457,"href":458},"en","\u002Fen\u002Fguides\u002Fprompting",{"loc":460,"href":461},"de","\u002Fde\u002Fratgeber\u002Fprompting",1789134650170]