--- title: "Brief vs Prompt: Give Your AI Agent a Defined Job." description: "Compare a brief vs prompt for recurring AI work. Define the job, allowed sources, decision rules, approval boundary, stopping point, and acceptance test." canonical: "https://scalewithsearch.com/articles/brief-vs-prompt" date: "2026-08-17" modified: "2026-09-19" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # A Brief Is Not a Prompt. You paste, "Write the client update in my voice." The draft sounds plausible, omits the current blocker, mentions unapproved work, and closes with a commitment nobody made. The prompt described a desired activity. It did not define the job. A task brief tells the agent what result to prepare, which records govern it, how to handle conflicts, what counts as complete, and where to stop. ## Prompts operate the model A prompt is the instruction sent to a model. It can set a role, provide context, request a format, or demonstrate examples. Model providers publish prompting guidance because instruction structure affects output. OpenAI's developer guidance covers message roles, instructions, examples, and reusable prompts. These are useful controls at the model interface. [OpenAI prompt engineering guide](https://developers.openai.com/api/docs/guides/prompt-engineering) The interface still needs a business specification behind it. "Summarize this account" leaves several questions unanswered: - Which account record is authoritative? - What date range matters? - Which facts are private? - Who will read the summary? - What decision should it support? - What should happen when records conflict? - May the agent send the result? A longer prompt can contain these answers. Once the instruction becomes recurring business infrastructure, it should live as a maintained brief rather than an improvised chat message. ## Briefs define the work A brief is model-independent. It should remain useful to a human operator, a new agent, or a replacement vendor. At minimum, it names: ```text job:: prepare weekly client update reader:: client account owner sources:: context.md, work-log.md, decisions.md, corrections.md source_order:: decisions.md overrides older work-log entries output:: weekly-update.md acceptance:: completed work, open blocker, next approved action must_not:: promise dates or scope not present in decisions.md stopping_point:: draft only receipt:: receipts/YYYY.MM.DD-weekly-update.md ``` The runtime prompt can then be short: "Run `task-brief.md` for the current week." The brief carries the durable job definition. The prompt initiates the run. ## Give each field one responsibility A useful brief separates six decisions that often get compressed into one long instruction. The job states the business result. The reader states who will use it. The sources state where facts may come from. The output contract states the file and required sections. The acceptance checks state what a reviewer can verify. The boundary states what the agent cannot do. Those decisions change at different speeds. The weekly input changes every run. The output structure may stay stable for months. The approval owner may change when a role changes. A reusable brief makes each change visible instead of burying all six inside revised prose. Use fields that a person can inspect without learning a prompt language: ```text job:: prepare a weekly account brief reader:: account owner current_input:: work-log/2026-W34.md allowed_sources:: account-context.md, decisions.md, corrections.md output_path:: drafts/2026-W34-account-brief.md required_sections:: completed work, open blocker, next approved action acceptance_checks:: source citations present, no unsupported date, no external send approval_owner:: account owner stopping_point:: draft and receipt written ``` Do not add a field because it sounds formal. Add it because a future operator needs to make a decision from it. If nobody can explain how `tone:: professional` will be checked, replace it with an approved example or a concrete language rule. ## Keep source precedence inside the brief The brief owns the rule that resolves competing sources for this job. The context packet may contain current records and an archive, but the job still needs to say which class of record governs. For example, an accepted decision file may override an earlier meeting transcript. A current account record may govern factual status while an approved sample governs layout only. Write those roles directly: ```text source_roles: decisions.md:: current approved choices account-context.md:: current account facts accepted-example.md:: structure only, facts not reusable archive/:: evidence only, never current by default ``` This rule prevents polished old material from winning because it is longer or easier to retrieve. If two current sources have the same authority and disagree, the brief should require a stop. The agent reports the conflict instead of choosing the version that makes completion easier. ## Put changeable facts outside the instruction Do not hard-code every price, client status, or deadline inside a reusable prompt. Put current facts in source files. Let the brief point to those files. This avoids editing the workflow instruction every time the business changes. For example, `task-brief.md` may say "Use the current approved offer from `offers.md`." The offer file can change through review while the job definition remains stable. This also makes failures easier to diagnose. If the output uses an old price, inspect retrieval and the offer record. If the output uses the wrong structure, inspect the brief or prompt adapter. ## Add an acceptance test Turn voice guidance into checks a reviewer can apply: - Use the section headings in `accepted-example.md`; its facts remain non-reusable. - Do not write a commitment such as "we will deliver Friday" unless `decisions.md` authorizes that date and scope. These rules make the intended professional tone observable without treating a polished sample as current policy. "Looks good" is weak acceptance criteria. A useful test names observable properties: - every status claim cites a named source file; - the draft contains the current blocker; - no date appears unless it exists in an approved source; - no external message is sent; - the receipt lists files read and files written. The agent can check some conditions. A human should review conditions that depend on relationship judgment or material business consequences. ## Inspect `task-brief.md` Take the prompt you reuse most. Move its business rules into `task-brief.md`. Then perform a replacement test: 1. Give the brief and its source files to a fresh agent session. 2. Do not provide the original chat thread. 3. Ask for the defined output and receipt. 4. Compare the result with the acceptance list. Pass means the job can be understood without your remembered explanation. Fail means the brief lacks a source, rule, or test. Repeat the test with a second model when the work matters. The outputs need not match word for word. They should obey the same facts and boundaries. ## Keep examples labeled Examples can improve output, but they need a status. Mark whether an example is approved, outdated, synthetic, or formatting-only. Otherwise an agent may copy an old claim because it appears in a polished sample. Use labels such as: ```text example_status:: formatting only facts_reusable:: no ``` This keeps demonstration material from becoming accidental policy. ## Separate the brief from the runtime adapter The same brief may be used by a person, a chat interface, an API workflow, or a different model. Each route can need a different adapter instruction. The business contract should not be rewritten for every route. Store vendor-specific details such as message roles, tool schemas, model identifiers, and response parsing beside the adapter. Keep job purpose, sources, acceptance, approval, and stopping in the task brief. That separation gives you a useful diagnostic. When two models receive the same brief and one omits a required section, inspect the adapter and model behavior. When both use an outdated account fact, inspect the source record or brief. The error has a smaller place to hide. ## Approval and stopping boundary The brief must state whether the agent may research, draft, edit files, or act externally. For client updates, proposals, emails, posts, bookings, purchases, deletions, deployments, and live record changes, default to preparation. The agent produces the exact proposed action and stops for approval. If the brief does not grant a capability, the agent does not infer it from urgency or convenience. ## Write the job once A prompt can get a good answer today. A brief lets the operation get a governed answer next week, from a fresh session, with a record of what happened. If several agents lack a shared definition of what to read and produce, write one job into a brief first. Preparing the [six-question brief](/work#prepare-your-six-question-brief) needs no purchase. [Send your brief](/work) ## Questions about A Brief Is Not a Prompt ### What exactly is a prompt? A prompt is the instruction sent to a model to set a role, provide context, request an output, or demonstrate examples. A task brief defines the recurring job, authoritative sources, source order, acceptance checks, approval boundary, and stopping point. ### How to actually write a prompt? A longer prompt can include business rules, but recurring infrastructure should place those rules in a maintained brief. The runtime prompt can then initiate the run while the brief remains inspectable by a human, a replacement model, or another operator. ### How would you phrase this for a brief to an external partner? A useful brief names the job, reader, allowed sources, output path, acceptance checks, approval owner, and stopping point. It also states what happens when sources conflict and whether the agent may only draft or may perform an external action. ## Save the visual summary Give each field one responsibility | Keep source precedence inside the brief | Put changeable facts outside the instruction | Add an acceptance test [Download the PNG](/infographics/brief-vs-prompt-1200x1500.png) ## Related: Owned Memory - [The Intake Is the Product](/articles/the-intake-is-the-product) - [Build an always-on AI automation stack in dependency order.](/articles/ai-automation-stack-explained) - [Clawdbot vs Claude Code: Computer Use or Terminal Work for a Business Job?](/articles/clawdbot-vs-claude-code-computer-terminal-work) ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```