Create a practical prompt pack for my work as a [SCIENTIFIC ROLE] in a [LAB TYPE OR FIELD]. My current project concerns [PROJECT CONTEXT], using these methods or instruments: [METHODS AND INSTRUMENTS]. The applicable safety, quality, and documentation requirements are: [LAB REQUIREMENTS]. My immediate priorities are [CURRENT TASKS]. Produce six copy-paste prompts, each 90–140 words, for: experiment planning, pre-run checklist, troubleshooting unexpected results, data-quality review, literature-reading notes, and lab-record or methods drafting. Each prompt must ask for the information needed to do that task, specify a useful output format, and distinguish observations from interpretations. Build sound laboratory practice into each prompt. Require controls, replicates, acceptance criteria, sample identity or chain-of-custody details where relevant, instrument or reagent lot traceability, and deviations from the protocol. For troubleshooting, require the model to rank plausible causes by evidence and propose low-risk diagnostic checks before changes to a validated procedure. For data review, require reporting of exclusions, missing data, outliers, and assumptions rather than silently cleaning them away. Do not provide operational instructions that override a validated SOP, biosafety requirement, clinical protocol, or qualified supervisor. Do not fabricate literature citations, experimental results, calibration status, or compliance records. Include a one-sentence safety or escalation reminder in every prompt where it is relevant. Before answering, check that all six prompts can be completed using real lab records and that none treats AI output as a substitute for required review or sign-off. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [SCIENTIFIC ROLE] | Enter your role, such as research assistant, analytical chemist, or molecular biologist. | Research associate |
| [LAB TYPE OR FIELD] | Describe the laboratory setting and scientific discipline. | Academic immunology laboratory |
| [PROJECT CONTEXT] | Briefly describe the study, assay, product, or research question. | Comparing cytokine response in stored PBMC samples after two stimulation conditions |
| [METHODS AND INSTRUMENTS] | List the relevant approved methods, instruments, and software. | Approved PBMC thawing SOP, ELISA, plate reader, R for analysis |
| [LAB REQUIREMENTS] | State applicable SOPs, safety levels, quality systems, and recordkeeping rules. | BSL-2 practices, sample IDs in ELN, reagent lot recording, PI review before protocol changes |
| [CURRENT TASKS] | List the work you need help organizing or reviewing now. | Plan the next plate, investigate variable control readings, and prepare a methods note |
How to use
- Replace the bracketed fields with your lab’s real context and approved requirements.
- Choose the prompt that matches the task, then paste in only the records and observations authorized for the AI tool.
- Check that proposed next steps remain within your SOP and document any deviation or decision in the required system.
- Follow up with: “Convert this troubleshooting output into a deviation-investigation outline, leaving all conclusions as pending.”
Variations
Experiment planner
Use this before a planned run to expose missing controls, dependencies, and acceptance criteria.
Help me plan an experiment within this approved protocol: [SOP OR METHOD SUMMARY]. The objective is [OBJECTIVE], samples are [SAMPLE DETAILS], and available resources are [CONSTRAINTS]. Produce a one-page planning checklist with hypothesis, variables, positive and negative controls, replicate rationale, sample randomization or blinding considerations, reagents and lot fields, instrument readiness checks, predefined acceptance criteria, and decision points. Identify assumptions and missing information. Do not alter the approved procedure or provide unsafe handling instructions; state when supervisor or QA approval is required.
Result triage
Use this after a run produces unexpected values, failed controls, or a confusing pattern.
Help triage unexpected laboratory results from [ASSAY OR METHOD]. Expected behavior was [EXPECTED RESULT]; observed data and run notes are [OBSERVATIONS]. Relevant controls, reagent lots, instrument checks, and deviations are [RUN RECORDS]. Return a ranked hypothesis table separating evidence, gaps, and non-destructive checks. Include sample identity, control performance, calculation, transcription, reagent, and instrument causes where applicable. State which results should be treated as invalid or pending review under [ACCEPTANCE CRITERIA]. Do not claim a root cause, change an SOP, or discard data without documented evidence and required oversight.
Methods draft
Use this when preparing a clear methods section or internal technical record from completed work.
Draft a factual [DOCUMENT TYPE] for [EXPERIMENT OR STUDY] using these contemporaneous records: [LAB RECORDS]. Follow this style guide or journal requirements: [STYLE REQUIREMENTS]. Produce sections for materials, sample handling, method, controls, data processing, deviations, and limitations. Use past tense for completed work and retain exact approved names, versions, units, and identifiers. Put missing details in a separate “information needed” list. Do not invent protocol steps, citations, approvals, instrument settings, or results. Check that another scientist could distinguish what was done from interpretation of the outcome.
Tips
- Treat AI-generated troubleshooting lists as hypotheses to test, not as explanations; preserve failed runs and control data until your review process decides otherwise.
- Prompt with the approved method version, acceptance criteria, and run record, because a generic method description cannot determine whether a deviation is material.
- Keep identifiable participant, patient, proprietary, or controlled data out of unapproved AI systems; use your institution’s approved tools and data-handling rules.
- For methods writing, give the AI your actual ELN entries and ask it to identify omissions rather than reconstructing a run from memory.
FAQ
Can AI help troubleshoot a failed assay?
It can organize hypotheses and diagnostic checks from your records, but it cannot validate a root cause or replace SOP, QA, safety, or supervisor decisions.
Can I use AI to write a methods section?
Yes, as a drafting aid from contemporaneous records. Check every parameter, reagent, version, and citation against the primary record before submission.
Should I upload raw lab data?
Only if your organization permits that tool and data type. Remove or avoid sensitive, regulated, identifiable, and proprietary information when required.