{"course":{"slug":"choosing-your-tools","title":"Choosing and Using the Tools","url":"https://addaly.com/learn/choosing-your-tools","level":"beginner","track":"work"},"lesson":{"slug":"from-chat-to-workflow","title":"From chatting to a workflow","url":"https://addaly.com/learn/choosing-your-tools/from-chat-to-workflow","position":68,"of":76,"minutes":8,"answer":"The people who get most from these tools have stopped starting from nothing: four or five saved instructions containing context, format, standards and one example they liked.","keyIdea":"The people who get most from these tools have stopped starting from nothing: four or five saved instructions containing context, format, standards and one example they liked, kept in their own plain text file rather than only inside a product's project feature."},"body":{"format":"markdown","text":"## What separates people who are fast\n\nWatch somebody who gets a great deal out of these tools and somebody who does not, and the difference is rarely prompting skill. It is that the fast one has **stopped starting from nothing**.\n\nThey have four or five tasks they do repeatedly, and for each one they have a saved instruction that already contains the context, the format, the standards and the examples. They paste the new material and get a usable answer on the first attempt. The slow one types a fresh description of the same job every time, gets a generic answer, and spends three turns steering it back.\n\nThe gap is not two minutes per task. It is that the first person's answer is good enough to use and the second person's is not.\n\n## What a saved instruction contains\n\nTake the last five times you asked for the same kind of thing, and write down what you had to explain every time:\n\n```\nROLE:     you are helping a two-person accountancy practice in Lagos\nCONTEXT:  clients are small traders; they read English but not jargon\nFORMAT:   short paragraphs, no bullet lists, under 200 words\nSTANDARDS: never state a tax figure without saying it must be checked\n          against the current schedule; never guess a deadline\nEXAMPLE:  <one reply you were happy with>\nTASK:     <the new thing goes here>\n```\n\nThe example is the part people leave out and the part that does most of the work. One instance of an answer you liked communicates more about tone, length and register than a paragraph describing it, and it is the cheapest quality improvement available anywhere in this course.\n\n## Where to keep it\n\nProducts offer projects, custom instructions, saved prompts, and named assistants. All of them are convenient and all of them live inside one company's product.\n\nKeep the master copy in a plain text file of your own, and paste it into whichever product's feature you are using. Then when you switch tools, or a feature is retired, or you want the same instruction on your local model, you already have it. This takes no extra effort (you were writing it anyway) and it is the difference between a workflow you own and one you rent.\n\n## The three levels, and when to move up\n\n**Level one: a saved prompt you paste.** Works everywhere, needs nothing, portable. Most people should stop here, and most never reach it.\n\n> Figure: What a saved instruction contains. Layers, top to bottom: ROLE (Who you are helping and at what scale. A two-person practice is not a company.); CONTEXT (Who reads the output, and what can be assumed about what they already know.); FORMAT (Length, structure, whether lists are wanted. Said once instead of corrected weekly.); STANDARDS (What must never be stated without a check. This is where your professional judgement lives.); EXAMPLE (One reply you were happy with. The part people leave out, and the part that does most of the work.); TASK (The new material, last: where the model attends most reliably, and after the cacheable prefix.). Keep the master copy in a plain text file of your own and paste it into whichever product's project feature you are using. You were writing it anyway, and it is the difference between a workflow you own and one you rent.\n\n**Level two: a project or custom instruction.** The product applies your context automatically. Convenient, one product only.\n\n**Level three: a script.** For anything you do more than a few times a day, or over more than a handful of items, a short script with the instruction inside it removes the copying and pasting entirely. The threshold is roughly: if you are doing the same thing twenty times, write the loop.\n\n## The habit worth building\n\nWhen an answer comes back good, ask one question: **will I want this again?**\n\nIf yes, spend ninety seconds saving what produced it: the instruction, the example, the correction that fixed it. That is the entire discipline. People who do this accumulate a set of tools shaped exactly to their work. People who do not have the same conversation every week for two years.\n\n## Where the instruction should live in the request\n\nOne detail that changes results more than its size suggests. Put the long stable material first and the specific task last. Reference documents, standards and examples at the top; the new item and the question at the bottom.\n\nThis helps for two separate reasons covered earlier in the course: models attend more reliably to the end of a long input, so the actual request lands where it is noticed, and prefix caching on paid APIs matches on the unchanging beginning, so the stable half is billed at a fraction. The same ordering is right whether you are paying by the token or not.\n\n## What not to systematise\n\nNot everything should become a workflow. Exploratory thinking, one-off questions, anything where you do not yet know what good looks like: these are better done loosely, and forcing them into a template produces confidently formatted answers to the wrong question.\n\nThe rule of thumb: systematise the tasks where you already know what a good answer looks like, and stay conversational where you do not. The first is production; the second is thinking, and they benefit from different handling."},"check":{"question":"Why does including one example of an answer you were happy with improve results more than describing the style you want?","options":[{"text":"Because the model treats examples as higher-priority instructions than prose, so they override conflicting guidance.","correct":false,"tempting":"Invents a priority ordering; examples work by showing the target rather than by outranking other instructions."},{"text":"Because the model can copy the example's structure directly, which removes the need for it to make any judgements.","correct":false,"tempting":"Overstates the mechanism into template-filling, when the model still adapts the shape to different material."},{"text":"Because examples are shorter than descriptions, leaving more of the context window for the task itself.","correct":false,"tempting":"Reverses the usual lengths, and a full example commonly costs more tokens than a sentence describing it."},{"text":"Because an example demonstrates tone, length and register at once, where a description only approximates them.","correct":true}]},"prev":{"slug":"specialist-tools","title":"Specialist and industry tools","url":"https://addaly.com/learn/choosing-your-tools/specialist-tools"},"next":{"slug":"portable-prompts","title":"Keeping your work portable","url":"https://addaly.com/learn/choosing-your-tools/portable-prompts"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'From chatting to a workflow', in Choosing and Using the Tools. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/choosing-your-tools/from-chat-to-workflow (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"From chatting to a workflow","courseTitle":"Choosing and Using the Tools","module":{"index":8,"slug":"living-with-it","title":"Living with it"},"answer":null,"keyIdea":"The people who get most from these tools have stopped starting from nothing: four or five saved instructions containing context, format, standards and one example they liked, kept in their own plain text file rather than only inside a product’s project feature.","hero":{"kind":"figure","figure":{"t":"figure","label":"Figure 8.1","figure":{"type":"stack","title":"What a saved instruction contains","layers":[{"label":"ROLE","detail":"Who you are helping and at what scale. A two-person practice is not a company."},{"label":"CONTEXT","detail":"Who reads the output, and what can be assumed about what they already know."},{"label":"FORMAT","detail":"Length, structure, whether lists are wanted. Said once instead of corrected weekly."},{"label":"STANDARDS","detail":"What must never be stated without a check. This is where your professional judgement lives."},{"label":"EXAMPLE","detail":"One reply you were happy with. The part people leave out, and the part that does most of the work."},{"label":"TASK","detail":"The new material, last: where the model attends most reliably, and after the cacheable prefix."}],"caption":"Keep the master copy in a plain text file of your own and paste it into whichever product’s project feature you are using. You were writing it anyway, and it is the difference between a workflow you own and one you rent."},"alt":"Figure: What a saved instruction contains. Layers, top to bottom: ROLE (Who you are helping and at what scale. A two-person practice is not a company.); CONTEXT (Who reads the output, and what can be assumed about what they already know.); FORMAT (Length, structure, whether lists are wanted. Said once instead of corrected weekly.); STANDARDS (What must never be stated without a check. This is where your professional judgement lives.); EXAMPLE (One reply you were happy with. The part people leave out, and the part that does most of the work.); TASK (The new material, last: where the model attends most reliably, and after the cacheable prefix.). Keep the master copy in a plain text file of your own and paste it into whichever product’s project feature you are using. You were writing it anyway, and it is the difference between a workflow you own and one you rent."}},"blocks":[{"t":"h2","id":"what-separates-people-who-are-fast","runs":[{"text":"What separates people who are fast"}]},{"t":"p","runs":[{"text":"Watch somebody who gets a great deal out of these tools and somebody who does not, and the difference is rarely prompting skill. It is that the fast one has "},{"text":"stopped starting from nothing","b":true},{"text":"."}]},{"t":"p","runs":[{"text":"They have four or five tasks they do repeatedly, and for each one they have a saved instruction that already contains the context, the format, the standards and the examples. They paste the new material and get a usable answer on the first attempt. The slow one types a fresh description of the same job every time, gets a generic answer, and spends three turns steering it back."}]},{"t":"p","runs":[{"text":"The gap is not two minutes per task. It is that the first person’s answer is good enough to use and the second person’s is not."}]},{"t":"h2","id":"what-a-saved-instruction-contains","runs":[{"text":"What a saved instruction contains"}]},{"t":"p","runs":[{"text":"Take the last five times you asked for the same kind of thing, and write down what you had to explain every time:"}]},{"t":"code","lang":"","label":"Text","text":"ROLE:     you are helping a two-person accountancy practice in Lagos\nCONTEXT:  clients are small traders; they read English but not jargon\nFORMAT:   short paragraphs, no bullet lists, under 200 words\nSTANDARDS: never state a tax figure without saying it must be checked\n          against the current schedule; never guess a deadline\nEXAMPLE:  <one reply you were happy with>\nTASK:     <the new thing goes here>"},{"t":"p","runs":[{"text":"The example is the part people leave out and the part that does most of the work. One instance of an answer you liked communicates more about tone, length and register than a paragraph describing it, and it is the cheapest quality improvement available anywhere in this course."}]},{"t":"h2","id":"where-to-keep-it","runs":[{"text":"Where to keep it"}]},{"t":"p","runs":[{"text":"Products offer projects, custom instructions, saved prompts, and named assistants. All of them are convenient and all of them live inside one company’s product."}]},{"t":"p","runs":[{"text":"Keep the master copy in a plain text file of your own, and paste it into whichever product’s feature you are using. Then when you switch tools, or a feature is retired, or you want the same instruction on your local model, you already have it. This takes no extra effort (you were writing it anyway) and it is the difference between a workflow you own and one you rent."}]},{"t":"h2","id":"the-three-levels-and-when-to-move-up","runs":[{"text":"The three levels, and when to move up"}]},{"t":"p","runs":[{"text":"Level one: a saved prompt you paste.","b":true},{"text":" Works everywhere, needs nothing, portable. Most people should stop here, and most never reach it."}]},{"t":"p","runs":[{"text":"Level two: a project or custom instruction.","b":true},{"text":" The product applies your context automatically. Convenient, one product only."}]},{"t":"p","runs":[{"text":"Level three: a script.","b":true},{"text":" For anything you do more than a few times a day, or over more than a handful of items, a short script with the instruction inside it removes the copying and pasting entirely. The threshold is roughly: if you are doing the same thing twenty times, write the loop."}]},{"t":"h2","id":"the-habit-worth-building","runs":[{"text":"The habit worth building"}]},{"t":"p","runs":[{"text":"When an answer comes back good, ask one question: "},{"text":"will I want this again?","b":true}]},{"t":"p","runs":[{"text":"If yes, spend ninety seconds saving what produced it: the instruction, the example, the correction that fixed it. That is the entire discipline. People who do this accumulate a set of tools shaped exactly to their work. People who do not have the same conversation every week for two years."}]},{"t":"h2","id":"where-the-instruction-should-live-in-the-request","runs":[{"text":"Where the instruction should live in the request"}]},{"t":"p","runs":[{"text":"One detail that changes results more than its size suggests. Put the long stable material first and the specific task last. Reference documents, standards and examples at the top; the new item and the question at the bottom."}]},{"t":"p","runs":[{"text":"This helps for two separate reasons covered earlier in the course: models attend more reliably to the end of a long input, so the actual request lands where it is noticed, and prefix caching on paid APIs matches on the unchanging beginning, so the stable half is billed at a fraction. The same ordering is right whether you are paying by the token or not."}]},{"t":"h2","id":"what-not-to-systematise","runs":[{"text":"What not to systematise"}]},{"t":"p","runs":[{"text":"Not everything should become a workflow. Exploratory thinking, one-off questions, anything where you do not yet know what good looks like: these are better done loosely, and forcing them into a template produces confidently formatted answers to the wrong question."}]},{"t":"p","runs":[{"text":"The rule of thumb: systematise the tasks where you already know what a good answer looks like, and stay conversational where you do not. The first is production; the second is thinking, and they benefit from different handling."}]}],"check":{"question":[{"text":"Why does including one example of an answer you were happy with improve results more than describing the style you want?"}],"options":[{"text":[{"text":"Because the model treats examples as higher-priority instructions than prose, so they override conflicting guidance."}],"correct":false,"tempting":[{"text":"Invents a priority ordering; examples work by showing the target rather than by outranking other instructions."}]},{"text":[{"text":"Because the model can copy the example’s structure directly, which removes the need for it to make any judgements."}],"correct":false,"tempting":[{"text":"Overstates the mechanism into template-filling, when the model still adapts the shape to different material."}]},{"text":[{"text":"Because examples are shorter than descriptions, leaving more of the context window for the task itself."}],"correct":false,"tempting":[{"text":"Reverses the usual lengths, and a full example commonly costs more tokens than a sentence describing it."}]},{"text":[{"text":"Because an example demonstrates tone, length and register at once, where a description only approximates them."}],"correct":true,"tempting":null}],"order":[2,3,0,1],"why":null},"prev":{"slug":"specialist-tools","title":"Specialist and industry tools","minutes":8},"next":{"slug":"portable-prompts","title":"Keeping your work portable","minutes":7}}}