Before hiring, see what AI can take on.
Before you hire another person to do the job as it exists today, find out which parts could work differently. Have AI identify promising pieces of the role, build small tools to try them, and show what still needs a person.
Try it now
Copy the prompt into your AI tool of choice, then answer its questions as they come.
Paste it into any AI chat to start. For most people, this is all you need.
Everything on this page in one paste: the prompt plus the full plan, every step and what to check. Use it for longer moves, to plan ahead, to share with a colleague, or if your AI can't open links.
How it goes
What the AI will do
- Investigate the work behind a proposed hire and identify tasks that might suit AI.
- Propose, build and test small tools on representative work, then compare quality, time and remaining human effort.
What to have ready
- A role, hiring request or rough description of work you need done.
- Available job descriptions, workload records and examples of the actual work; the AI searches relevant connected records first.
Where you come in
- Clarify what good work looks like and choose the tools worth testing.
- Try or review the results where your expertise is needed, then decide what they mean for the role.
What you'll get
Working trial tools, evidence of what they can and cannot do, and a revised staffing brief grounded in those trials.
You're done when
Working trial tools, evidence of what they can and cannot do, and a revised staffing brief grounded in those trials.
These are general instructions, not a description of this user's particular AI system, tools, access, or circumstances. Adapt the method to the user's goal and the capabilities you can actually use, while preserving the activity's essential sequence, review points, and intended result. ### Start with the activity and the user's context Read the complete activity and guide the user through it. Use the activity prompt yourself; the user does not need to paste it again. Keep the responsibilities of the user, the AI, and any colleague clear. Begin with what the user has already supplied. Establish the intended result and resolve missing context through focused questions when needed. Work within the activity's stated scope, permissions, required confirmations, and stopping points, including any deliberate order for introducing information. Make progress with the available information and reasonable, clearly identified assumptions where the activity allows them. ### Check the capabilities needed for this work Inspect relevant available tools, connectors, and data. Use small, relevant, non-destructive checks to establish whether necessary access or a capability actually works. Naming a product does not establish access to it. Let this activity determine what needs checking; do not impose a universal preliminary questionnaire or a full system audit. When useful access is missing: - Explain what you cannot reach and how that material would help. - Investigate practical setup, connection, export, or other access alternatives before asking the user to gather material or troubleshoot. - If progress depends on access, request only the specific user action or permission needed to enable it or provide the material another way. Continue useful independent work while awaiting the response. - Offer to walk the user through an export or upload. Welcome files, links, copied text, rough notes, or an interview as supported by your system and the activity, and use the best available route. If the missing access is optional, continue with what is available. ### Ask focused questions and respect the answers Ask questions only when necessary to complete the work well or satisfy the activity's required review points. Build on answers already supplied. Ask one question at a time, or a small group of closely related questions. Wait on work that depends on the answer, and continue useful work that does not. Explain unfamiliar language with examples where helpful. Confirm answers or decisions you supply on the user's behalf before treating them as the user's final answers or decisions. This does not require separate approval of every sourced finding; follow the activity's specific review requirements. ### Keep evidence and independent contributions distinct Distinguish evidence, inference, and missing information. Check your work against the source material. Use independent agents when your system actually supports delegation and independent investigation or review would improve the work: - Give each agent a bounded brief and the context appropriate to its role. - Keep discovery independent until synthesis where the activity requires it. Preserve any requirement for the user or colleagues to contribute before seeing other explanations. - Follow the activity's stated fallback and any required confirmation when independent agents or fresh contexts are unavailable. Preserve reviewer isolation where required; describe shared-context or sequential passes accurately. ### Produce, check, and revise the result Do the substantive investigation and production. Carry the activity through to the most complete, usable result within its scope, rather than ending with a plan or handing the main work back to the user. When missing information does not prevent useful work, proceed with clearly identified assumptions where the activity permits them. Check the result against the activity's completion criteria. At each substantive delivery, state what is complete and any material unfinished work, assumptions, helpful missing information, and remaining uncertainty. Explain how these affect what the user can use, and help the user revise the result. Preserve the activity's stopping point. Do not turn a rehearsal, investigation, or test-design activity into execution that the user has not authorized. ### Find the work behind the proposed hire Help me test what AI could take on before hiring for this role. Identify important parts of the job that might be AI-able, propose tools you can build to test those ideas, build the tools I choose and test them on actual representative work. Use the evidence to clarify the remaining human role. Do not stop at a task list or a recommendation to buy software. Start from the role, hiring request or rough description I provide. Inspect relevant available connections and search requisitions, job descriptions, workload records, project deliverables, process documents, tickets, calendars and managers’ email or chat discussions, including relevant WhatsApp exchanges where accessible. Use the role, team and project names to find why the hire is needed, what work arrives, who relies on it, where it stalls and what quality or judgment matters. Distinguish proposed duties from work people actually do. Investigate usable connection or export routes before asking me to assemble a packet. If source access is limited, work from a small set of representative examples or interview me one question at a time. Label estimates and unobserved work. Ask about the purpose or a material missing constraint, then wait before designing around an answer you do not have. ### Identify promising pieces and propose tools Break the role into meaningful outputs and workflows: what starts the work, its inputs, frequency or volume, ordinary cases, exceptions, required judgment, quality standards and recipients. Use recipients’ documented requests and feedback to understand what makes an output useful. Keep coordination, exception handling and accountability visible alongside production tasks. Look for consequential pieces where an AI-assisted tool might reduce work or create a better result. Examples might include preparing a source-linked customer brief, checking a recurring report for discrepancies, or turning incoming requests into complete draft work packets. Treat these as examples, not a mandatory tool set. For each promising candidate, propose a small tool with a concrete input, output, user, feasible implementation and representative success test. Explain what it would automate or assist, what the person would retain, and what the experiment would establish. Inspect your actual coding, agent-building and integration capabilities; use current official documentation when a capability claim matters. Prefer the company’s available tools where they serve the experiment. Offer a manageable shortlist and recommend where to start. Ask me which tools to build and resolve consequential quality or scope choices. A tool that only works on a polished ideal input is not a fair test of the role. ### Establish a comparison and build the tools Select representative cases from the work record, including an ordinary case, a difficult case and a common exception where available. Preserve some unseen cases for testing after development. Establish the expected output or evaluation criteria from prior accepted work and the relevant person’s judgment. Ask me or an appropriate reviewer to resolve standards the record cannot establish. Do not invent a human performance baseline. Record available baseline time, quality, cost and human effort, distinguishing measured values from estimates. Include preparation, checking, correction and supervision in the comparison, not just AI generation time. Build the selected tools using available capabilities. Create the actual prompts, workflow, interface and connections needed for the agreed trial. If several tools can be built independently, assign bounded tasks with the tool brief, permitted sources, expected output and success criteria; integrate dependent pieces centrally. Keep a coverage record and use successive waves when agent capacity is limited. Without agents, use separate manageable implementation and checking passes. Run a small development trial and correct failures. If a connection is unavailable, investigate supported alternatives and ask for the specific access action only when needed. A tool using an export can still test the agreed task, but identify the manual transfer it requires. A mock response or interface without the working core is not the completed trial tool. ### Test the actual work and account for what remains Run the held-back cases through the working tools. Record inputs, actual outputs, failures, corrections, elapsed time where measurable and human intervention. Check source fidelity, completeness and fitness for the recipient’s real use. Include malformed or missing inputs where they occur in this work. Do not count AI-produced answers as correct simply because another AI finds them persuasive. Use an independent checker when available, giving it the actual outputs, source material and quality criteria without the builder’s defense. Request domain judgment from me only where it is needed. Compare the resulting work with the baseline and distinguish independent checks, your own checks and my reported experience. Repair material defects and rerun affected cases. If the tool cannot meet the agreed standard, report the evidence rather than extending the project indefinitely or declaring the role replaceable. Estimate the human work left: handling exceptions, reviewing outputs, managing relationships, making decisions, maintaining the tool and work the experiment did not cover. Avoid turning a percentage of tasks into a percentage of a person without considering task frequency, dependencies and time. A brief trial establishes only its tested conditions. ### Deliver usable tools and a staffing decision brief Give me the working tools, source/configuration where applicable, short usage instructions, test cases and results. State precisely what is installed or connected and what remains incomplete. Summarize the role as tested work AI can take on, work AI can assist and work still requiring people or further investigation. Write a revised staffing brief that reflects those findings, with practical options such as changing the role’s emphasis, hiring for a different capability, or retaining the original role because the trials did not solve the need. Preserve the hiring choice for me. Complete the activity with the tools and evidence; a staffing opinion without built and tested tools is an intermediate result. These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.
View full AI instructions
Before hiring, see what AI can take on.
What the AI will do
- Investigate the work behind a proposed hire and identify tasks that might suit AI.
- Propose, build and test small tools on representative work, then compare quality, time and remaining human effort.
What you can provide
- A role, hiring request or rough description of work you need done.
- Available job descriptions, workload records and examples of the actual work; the AI searches relevant connected records first.
Where you'll take part
- Clarify what good work looks like and choose the tools worth testing.
- Try or review the results where your expertise is needed, then decide what they mean for the role.
What you'll get
Working trial tools, evidence of what they can and cannot do, and a revised staffing brief grounded in those trials.
Activity context
Before you hire another person to do the job as it exists today, find out which parts could work differently. Have AI identify promising pieces of the role, build small tools to try them, and show what still needs a person.
For the user
Start with a vacancy, a proposed role or work that is piling up. The AI investigates the need and suggests a few small tools to build. You choose a useful scope and judge the work that needs your expertise; the AI carries the building, testing and comparison.
Guidance for the AI
These are general instructions, not a description of this user's particular AI system, tools, access, or circumstances. Adapt the method to the user's goal and the capabilities you can actually use, while preserving the activity's essential sequence, review points, and intended result.
Start with the activity and the user's context
Read the complete activity and guide the user through it. Use the activity prompt yourself; the user does not need to paste it again. Keep the responsibilities of the user, the AI, and any colleague clear.
Begin with what the user has already supplied. Establish the intended result and resolve missing context through focused questions when needed. Work within the activity's stated scope, permissions, required confirmations, and stopping points, including any deliberate order for introducing information. Make progress with the available information and reasonable, clearly identified assumptions where the activity allows them.
Check the capabilities needed for this work
Inspect relevant available tools, connectors, and data. Use small, relevant, non-destructive checks to establish whether necessary access or a capability actually works. Naming a product does not establish access to it. Let this activity determine what needs checking; do not impose a universal preliminary questionnaire or a full system audit.
When useful access is missing:
- Explain what you cannot reach and how that material would help.
- Investigate practical setup, connection, export, or other access alternatives before asking the user to gather material or troubleshoot.
- If progress depends on access, request only the specific user action or permission needed to enable it or provide the material another way. Continue useful independent work while awaiting the response.
- Offer to walk the user through an export or upload. Welcome files, links, copied text, rough notes, or an interview as supported by your system and the activity, and use the best available route. If the missing access is optional, continue with what is available.
Ask focused questions and respect the answers
Ask questions only when necessary to complete the work well or satisfy the activity's required review points. Build on answers already supplied. Ask one question at a time, or a small group of closely related questions. Wait on work that depends on the answer, and continue useful work that does not. Explain unfamiliar language with examples where helpful.
Confirm answers or decisions you supply on the user's behalf before treating them as the user's final answers or decisions. This does not require separate approval of every sourced finding; follow the activity's specific review requirements.
Keep evidence and independent contributions distinct
Distinguish evidence, inference, and missing information. Check your work against the source material.
Use independent agents when your system actually supports delegation and independent investigation or review would improve the work:
- Give each agent a bounded brief and the context appropriate to its role.
- Keep discovery independent until synthesis where the activity requires it. Preserve any requirement for the user or colleagues to contribute before seeing other explanations.
- Follow the activity's stated fallback and any required confirmation when independent agents or fresh contexts are unavailable. Preserve reviewer isolation where required; describe shared-context or sequential passes accurately.
Produce, check, and revise the result
Do the substantive investigation and production. Carry the activity through to the most complete, usable result within its scope, rather than ending with a plan or handing the main work back to the user. When missing information does not prevent useful work, proceed with clearly identified assumptions where the activity permits them.
Check the result against the activity's completion criteria. At each substantive delivery, state what is complete and any material unfinished work, assumptions, helpful missing information, and remaining uncertainty. Explain how these affect what the user can use, and help the user revise the result.
Preserve the activity's stopping point. Do not turn a rehearsal, investigation, or test-design activity into execution that the user has not authorized.
Your AI prompt
Find the work behind the proposed hire
Help me test what AI could take on before hiring for this role. Identify important parts of the job that might be AI-able, propose tools you can build to test those ideas, build the tools I choose and test them on actual representative work. Use the evidence to clarify the remaining human role. Do not stop at a task list or a recommendation to buy software.
Start from the role, hiring request or rough description I provide. Inspect relevant available connections and search requisitions, job descriptions, workload records, project deliverables, process documents, tickets, calendars and managers’ email or chat discussions, including relevant WhatsApp exchanges where accessible. Use the role, team and project names to find why the hire is needed, what work arrives, who relies on it, where it stalls and what quality or judgment matters. Distinguish proposed duties from work people actually do.
Investigate usable connection or export routes before asking me to assemble a packet. If source access is limited, work from a small set of representative examples or interview me one question at a time. Label estimates and unobserved work. Ask about the purpose or a material missing constraint, then wait before designing around an answer you do not have.
Identify promising pieces and propose tools
Break the role into meaningful outputs and workflows: what starts the work, its inputs, frequency or volume, ordinary cases, exceptions, required judgment, quality standards and recipients. Use recipients’ documented requests and feedback to understand what makes an output useful. Keep coordination, exception handling and accountability visible alongside production tasks.
Look for consequential pieces where an AI-assisted tool might reduce work or create a better result. Examples might include preparing a source-linked customer brief, checking a recurring report for discrepancies, or turning incoming requests into complete draft work packets. Treat these as examples, not a mandatory tool set.
For each promising candidate, propose a small tool with a concrete input, output, user, feasible implementation and representative success test. Explain what it would automate or assist, what the person would retain, and what the experiment would establish. Inspect your actual coding, agent-building and integration capabilities; use current official documentation when a capability claim matters. Prefer the company’s available tools where they serve the experiment.
Offer a manageable shortlist and recommend where to start. Ask me which tools to build and resolve consequential quality or scope choices. A tool that only works on a polished ideal input is not a fair test of the role.
Establish a comparison and build the tools
Select representative cases from the work record, including an ordinary case, a difficult case and a common exception where available. Preserve some unseen cases for testing after development. Establish the expected output or evaluation criteria from prior accepted work and the relevant person’s judgment. Ask me or an appropriate reviewer to resolve standards the record cannot establish. Do not invent a human performance baseline.
Record available baseline time, quality, cost and human effort, distinguishing measured values from estimates. Include preparation, checking, correction and supervision in the comparison, not just AI generation time.
Build the selected tools using available capabilities. Create the actual prompts, workflow, interface and connections needed for the agreed trial. If several tools can be built independently, assign bounded tasks with the tool brief, permitted sources, expected output and success criteria; integrate dependent pieces centrally. Keep a coverage record and use successive waves when agent capacity is limited. Without agents, use separate manageable implementation and checking passes.
Run a small development trial and correct failures. If a connection is unavailable, investigate supported alternatives and ask for the specific access action only when needed. A tool using an export can still test the agreed task, but identify the manual transfer it requires. A mock response or interface without the working core is not the completed trial tool.
Test the actual work and account for what remains
Run the held-back cases through the working tools. Record inputs, actual outputs, failures, corrections, elapsed time where measurable and human intervention. Check source fidelity, completeness and fitness for the recipient’s real use. Include malformed or missing inputs where they occur in this work. Do not count AI-produced answers as correct simply because another AI finds them persuasive.
Use an independent checker when available, giving it the actual outputs, source material and quality criteria without the builder’s defense. Request domain judgment from me only where it is needed. Compare the resulting work with the baseline and distinguish independent checks, your own checks and my reported experience. Repair material defects and rerun affected cases. If the tool cannot meet the agreed standard, report the evidence rather than extending the project indefinitely or declaring the role replaceable.
Estimate the human work left: handling exceptions, reviewing outputs, managing relationships, making decisions, maintaining the tool and work the experiment did not cover. Avoid turning a percentage of tasks into a percentage of a person without considering task frequency, dependencies and time. A brief trial establishes only its tested conditions.
Deliver usable tools and a staffing decision brief
Give me the working tools, source/configuration where applicable, short usage instructions, test cases and results. State precisely what is installed or connected and what remains incomplete. Summarize the role as tested work AI can take on, work AI can assist and work still requiring people or further investigation.
Write a revised staffing brief that reflects those findings, with practical options such as changing the role’s emphasis, hiring for a different capability, or retaining the original role because the trials did not solve the need. Preserve the hiring choice for me. Complete the activity with the tools and evidence; a staffing opinion without built and tested tools is an intermediate result.
Done when
Working trial tools, evidence of what they can and cannot do, and a revised staffing brief grounded in those trials.
These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.
These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.