25 Best ChatGPT Prompts for Business in 2026

ChatGPT is most useful at work when you give it a real business task, the source material a colleague would need, and a clear standard for the finished output.

The 25 prompts below are built around that idea. They turn customer notes, spreadsheets, drafts and decisions into useful artifacts: a market brief, a campaign plan, a sales follow-up, an SOP, a cash-flow scenario or an interview scorecard.

Each prompt has fillable brackets. Replace them with real information, attach the relevant files where possible, and treat the response as work to review rather than truth delivered from a very confident cloud.

The best ChatGPT prompts for business at a glance

How to get better results from these prompts

OpenAI’s workplace prompting guidance recommends defining the task, supplying useful context and describing the ideal output. Its current prompt engineering guidance also recommends clear instructions, relevant examples and explicit output formats.

You do not need every detail you know about the business. Include the information that could change the answer:

• What the business sells and to whom
• The decision or deliverable you need
• Relevant evidence, such as analytics, call notes, policies or examples
• Constraints involving budget, time, brand, systems or approvals
• The exact format you want back
• What ChatGPT must not invent or assume

For recurring work, save this context block and paste it before a prompt:

BUSINESS CONTEXT
Company: [name and one-sentence description]
Offer: [products or services]
Customers: [main customer segments]
Business model: [how the company makes money]
Current priorities: [top goals]
Constraints: [budget, capacity, policies, tools or deadlines]
Voice: [how the company communicates]
Non-negotiables: [claims, topics, actions or data the output must avoid]

When prompts are not enough

If the task depends on many files or frequently changing data, stop copying fragments into each prompt. Give your AI assistant controlled access to the relevant source through file uploads, connected apps, plugins or an approved MCP integration. It can then work from the same CRM, drive, spreadsheet or data warehouse your team uses.

ChatGPT Work can work across connected tools and files. In the ChatGPT desktop app, it can also use selected local files and desktop apps with your permission. Other AI assistants offer similar integrations.

Grant only the access required for the task, keep consequential actions behind human approval, and verify the result before it reaches customers or business systems.

Use approved or anonymized data. OpenAI states that it does not train on inputs and outputs from ChatGPT Business, Enterprise and its API Platform by default. Consumer ChatGPT accounts have separate data controls, including an option to turn off model training. Your company’s own privacy, security and compliance rules still apply.

Strategy and planning prompts

1. Assess a business opportunity

Use this before committing money or months of work to a new offer, market or product idea. It forces ChatGPT to separate evidence from attractive speculation.

Create an evidence-based opportunity brief for this business idea.

Business idea: [describe the product, service or expansion]
Target customer: [who would buy it]
Problem: [problem it is meant to solve]
Revenue model: [how it would make money]
Available evidence: [customer interviews, search data, sales requests, competitor notes, market reports or other evidence]
Constraints: [budget, team, deadline, geography, regulation or technical limits]

Use only the evidence I provide and current sources you can cite. Do not invent market size, customer demand or competitor facts. If essential information is missing, ask up to 5 questions in one batch before continuing.

Return:
1. A 5-sentence opportunity summary
2. Evidence for and against the idea
3. The 5 assumptions most likely to break it
4. A simple competitor and substitute map
5. The cheapest test for each critical assumption
6. Clear success and stop criteria for a 30-day validation sprint
7. A recommendation: test, revise, defer or reject

Label every unsupported point as an assumption. End with the next 3 actions, in order.

Review the recommendation against real customer behavior. A neat brief cannot rescue an idea nobody wants.

Related resource: use GPTBot’s niche opportunity analyzer.

2. Define an ideal customer profile from evidence

This prompt builds an ideal customer profile from actual wins, losses and customer language. It avoids the usual fictional persona who somehow loves podcasts, sustainability and premium software.

Build an ideal customer profile using the evidence below.

Business: [what we sell and how it is priced]
Customer evidence: [paste customer list, CRM export, interview notes, reviews, won/lost deal notes or support themes]
Current hypothesis: [who we think the best customer is]
Commercial goal: [more qualified leads, higher retention, larger deals or another goal]

Analyze only the supplied evidence. Do not invent demographics, budgets, job titles or buying behavior. Mark weak conclusions clearly.

Return:
1. The strongest customer segment, with supporting evidence
2. Jobs they are trying to complete
3. Pain points in their own language
4. Buying triggers and urgency signals
5. Common objections and reasons not to buy
6. Minimum qualification criteria
7. Disqualifying criteria
8. A concise ICP statement the sales and marketing teams can use
9. Missing evidence and 5 research questions

Include a table mapping each conclusion to the evidence that supports it.

If the source data mixes very different customer types, run this once per segment rather than asking for one mushy average.

3. Analyze competitor positioning

Use this to understand how competitors present themselves, where they overlap and which claims are still unoccupied.

Analyze the positioning of these competitors.

Our company: [description, audience, offer and current positioning]
Competitors: [names and URLs]
Evidence: [paste homepage copy, pricing notes, reviews, sales feedback or product details]
Target market: [segment and geography]

If you have web access, use current first-party pages and cite every factual claim with its URL and access date. If you do not have access, use only the material I provide. Do not infer customer counts, revenue, market share or product capabilities.

Create:
1. A comparison table covering target customer, category, core promise, proof, pricing position, CTA and notable limitations
2. Repeated claims that make the market sound interchangeable
3. Customer needs each competitor appears to prioritize
4. Credible positioning gaps we could investigate
5. 3 differentiated positioning directions for us
6. Evidence needed before choosing a direction

For each proposed direction, state who it is for, the promise, proof required, tradeoffs and why competitors would struggle to copy it. Finish with a recommendation, not a slogan.

Use the findings to guide customer research. A competitor’s homepage tells you what it claims, not necessarily why customers buy.

4. Compare strategic options with a decision matrix

This prompt helps a team compare options using agreed criteria rather than whichever argument had the nicest slide.

Build a decision matrix for the choice below.

Decision: [what we need to decide]
Options: [option A, option B, option C]
Decision deadline: [date]
Success criteria: [criteria such as revenue, learning, speed, strategic fit, customer value or reversibility]
Constraints: [budget, people, time, technical limits or policy]
Evidence: [costs, estimates, research, customer data and known dependencies]

Before scoring, identify ambiguous criteria and ask up to 5 questions if needed.

Then:
1. Define each criterion so it can be scored consistently
2. Suggest percentage weights totaling 100 and explain them briefly
3. Score every option from 1 to 5 using only supplied evidence
4. Show the weighted calculation in a table
5. Separate facts, estimates and assumptions
6. Run a sensitivity check showing what changes if the 2 largest weights move by 10 percentage points
7. Identify risks, dependencies, reversibility and information gaps
8. Recommend an option and the evidence that would change the recommendation

Do not hide uncertainty behind precise scores. If evidence is too weak, say so.

Check the arithmetic and challenge the weights before using the result in a real decision.

5. Run a pre-mortem on an important plan

A pre-mortem assumes the plan failed and works backward. Use it before a launch, hire, migration, campaign or expansion.

Pressure-test this plan as if it failed badly 6 months from now.

Plan: [paste the plan or describe the decision]
Desired outcome: [what success means]
Timeline: [dates and milestones]
Resources: [budget, people, tools and partners]
Known concerns: [current doubts or constraints]
Evidence: [research, forecasts or customer input]

Return:
1. The 10 most plausible failure causes, ranked by likelihood and impact
2. Early warning signs for each cause
3. Assumptions the plan depends on
4. Which failures are preventable, transferable, acceptable or fatal
5. Mitigation actions with owners and deadlines
6. A smaller or more reversible version of the plan
7. A go, revise or stop recommendation

Argue against the plan fairly. Do not manufacture obscure risks to fill the list. Distinguish risks supported by evidence from informed speculation.

Run this with the people responsible for delivery. They usually know where the bodies are scheduled to be buried.

Marketing and content prompts

6. Build a focused 90-day marketing plan

This turns a commercial goal into a short plan with priorities, owners and measurement.

Create a 90-day marketing plan for this business.

Business and offer: [description]
Audience: [customer segment]
Commercial goal: [pipeline, sales, retention, launch or another outcome]
Baseline: [current traffic, leads, conversion, revenue and active channels]
Budget: [amount]
Team capacity: [people and hours]
Existing assets: [content, list, case studies, partnerships or tools]
Constraints: [brand, market, regulation, deadline or capabilities]

Do not assume every channel is necessary. Prioritize the few activities most likely to affect the commercial goal.

Return:
1. The target audience and buying situation
2. One measurable objective and supporting indicators
3. The core message and proof required
4. A ranked channel plan with rationale
5. A week-by-week execution calendar
6. Owners, dependencies and budget by activity
7. Experiments with hypotheses, success thresholds and stop rules
8. A simple weekly dashboard
9. What not to do during these 90 days

Flag missing baseline data. Do not invent benchmarks or expected returns.

For a longer planning workflow, use GPTBot’s content marketing strategy prompt.

7. Extract a practical brand voice guide

Use approved writing samples to document how the business actually sounds, not how an AI thinks a “bold but approachable” brand should sound.

Create a practical brand voice guide from the approved samples below.

Business: [description]
Audience: [who reads the content]
Channels: [website, email, sales, support, social or others]
Approved samples: [paste 3 to 10 examples]
Samples we dislike: [paste examples and explain why]
Required terminology: [product names, technical terms or legal wording]

Analyze observable patterns only. Do not assign vague personality traits unless the samples support them.

Return:
1. A 100-word voice summary
2. Sentence length, vocabulary, rhythm and paragraph patterns
3. How the voice changes by channel
4. Words and constructions to use
5. Words, clichés and habits to avoid
6. Formatting and punctuation rules
7. 5 before-and-after rewrites
8. A final checklist a writer can use before publishing

Quote short examples from the supplied samples to support each rule. Preserve factual meaning in every rewrite.

Save the result alongside real approved examples. The examples will teach the model more than a page of adjectives.

Related resource: analyze a writing style with this GPTBot prompt.

8. Mine customer language for messaging

Use call transcripts, reviews, surveys and support tickets to find the words customers use for their problem and desired outcome.

Analyze this customer research for useful messaging language.

Offer: [product or service]
Target segment: [customer group]
Source material: [paste reviews, transcripts, survey responses or tickets]
Goal: [homepage messaging, sales enablement, ads, onboarding or product research]

Use only the supplied text. Do not merge quotations, clean up grammar inside quotes or invent customer statements. Remove personal information from the output.

Return:
1. Repeated problems, desired outcomes, triggers, objections and alternatives
2. A frequency count for each theme
3. Exact quotations for the strongest themes, labeled by source
4. Differences between customer segments
5. Claims we can support and claims that need more evidence
6. 10 messaging angles using customer language without copying full quotations
7. 5 follow-up research questions

Separate what customers explicitly said from your interpretation. Note when the sample is too small or biased for a strong conclusion.

This works better with 20 real comments than a fictional persona generated from nothing.

9. Create an SEO content strategy from business priorities

This prompt connects search topics to the offer, customer journey and existing content. It should produce a roadmap, not 100 keywords arranged by vibes.

Create an SEO content strategy for this business.

Business: [description]
Offer and margins: [products or services to prioritize]
Audience: [customer segments]
Markets and languages: [locations]
Website: [URL]
Existing content: [URLs or export]
Keyword data: [paste data from Search Console, Semrush, Ahrefs or another source]
Competitors: [URLs]
Commercial goals: [qualified leads, trials, purchases or retention]
Constraints: [expert availability, publishing capacity, legal review or seasonality]

Use the supplied keyword data for volume and difficulty. Do not invent metrics. Group keywords by shared search intent, not just similar words.

Return:
1. Current content gaps and cannibalization risks
2. 3 to 5 topic clusters tied to business goals
3. A pillar page and supporting articles for each cluster
4. Primary query, intent, funnel stage and internal links for every page
5. A prioritized 12-week editorial calendar
6. Evidence or first-hand experience each article needs
7. Conversion action for each article
8. Pages to update, consolidate or leave alone

Rank opportunities by business relevance, realistic competitiveness and ability to create a genuinely better answer.

GPTBot also has a dedicated SEO content roadmap prompt.

10. Audit and rewrite a landing page

Use this when a page gets traffic but the offer, proof or next step is unclear.

Audit this landing page for conversion clarity, then rewrite only the weak sections.

Page copy or URL: [paste copy or URL]
Traffic source: [search, ads, email, referral or direct]
Target customer: [specific segment]
Offer: [what they receive]
Primary conversion: [demo, purchase, trial, call or form]
Available proof: [case studies, testimonials, data, guarantees or credentials]
Constraints: [brand voice, legal claims, page layout or required wording]

Evaluate:
- Message match with the traffic source
- Clarity of audience, problem, offer and outcome
- Strength and placement of proof
- Objections left unanswered
- Friction in the CTA and form
- Unsupported or vague claims
- Mobile scanning and heading structure

Return:
1. The 5 highest-impact problems, ranked
2. Evidence from the current copy for each problem
3. A revised message hierarchy
4. New copy for the hero, benefit section, proof section, objections and CTA
5. 3 testable hypotheses with one variable per test
6. Claims or proof that must be supplied before publishing

Do not invent testimonials, urgency, results, customers or product capabilities. Keep strong sections unchanged.

Use the GPTBot product-page conversion prompt for ecommerce pages.

Sales and customer communication prompts

11. Prepare a discovery call

This creates a call plan around what is still unknown rather than a rigid script the prospect has to survive.

Prepare a discovery call plan for this prospect.

Our offer: [description]
Ideal customer: [ICP]
Prospect: [company, role and known context]
Lead source: [how they found us]
Known needs: [facts from the inquiry or previous conversation]
Unknowns: [what we still need to learn]
Call length: [minutes]
Qualification criteria: [budget, urgency, authority, problem fit or other criteria]

Use only verified prospect facts. Label assumptions and do not manufacture personalization.

Return:
1. A 2-sentence call objective
2. A short opening and agenda
3. 8 open questions ordered naturally
4. Follow-up questions for likely answer paths
5. Signals of fit, risk and disqualification
6. Relevant proof or examples to prepare
7. A concise way to propose the next step
8. A note template for the CRM

Design the plan so the prospect speaks for most of the call. Avoid manipulative questions and premature pitching.

12. Create a sourced prospect research brief

Use this before contacting a high-value account. It keeps research factual and gives sales a reason to care about each finding.

Create a pre-call research brief for this prospect.

Prospect company: [name and URL]
Contact: [name, role and public profile]
Our offer: [description]
Ideal customer criteria: [criteria]
Research date: [date]
Approved sources: [company site, filings, job posts, interviews, public profiles or news]

Use current public sources and cite every factual claim with a URL and date. Prefer first-party sources. Do not infer private priorities, budget, technology or pain from a job title alone.

Return:
1. Company summary relevant to our offer
2. Verified recent changes or triggers
3. Relevant strategic priorities stated by the company
4. Evidence of fit and evidence against fit
5. Likely stakeholders, clearly labeled as hypotheses
6. 5 questions to validate needs on the call
7. 3 outreach angles tied to verified facts
8. Facts that remain unknown

Keep the brief under 700 words. Exclude trivia that does not change the sales approach.

13. Write cold outreach based on verified facts

This produces short outreach without fake familiarity, inflated claims or the phrase “I noticed.” Civilization survives another day.

Write 3 cold outreach email variants for this prospect.

Recipient: [name, role and company]
Verified facts we may reference: [facts with sources]
Relevant problem: [problem our offer addresses]
Our offer: [one sentence]
Proof we can support: [case study, result, experience or none]
Desired next step: [reply, short call or permission to send information]
Tone: [direct, friendly, technical or another style]

Requirements:
- Use only the verified facts and approved proof
- 80 to 120 words per email
- One idea and one CTA per email
- Make each variant use a different angle
- Do not use flattery, fake urgency or unsupported personalization
- Avoid "I noticed," "just checking in," "hope you're well," and similar filler
- Do not pretend we have an existing relationship

For each variant, include a subject line of 3 to 6 words and explain in one sentence why the angle may be relevant. Then flag any claim that needs approval.

See GPTBot’s cold outreach specialist prompt for another starting point.

14. Draft a sales proposal from discovery notes

This turns approved notes into a proposal without quietly inventing scope, timing or price.

Draft a sales proposal using the discovery information below.

Prospect: [company and stakeholders]
Discovery notes: [paste notes]
Their stated goals: [goals]
Current situation and impact: [facts]
Proposed solution: [approved scope]
Deliverables: [list]
Exclusions: [list]
Timeline: [approved dates or placeholder]
Price and payment terms: [approved figures or placeholder]
Proof: [relevant approved examples]
Decision process: [known steps]

Use only the supplied information. If a requirement, price, deadline or capability is missing, insert [CONFIRM] instead of guessing.

Structure the proposal as:
1. Executive summary
2. Current situation
3. Objectives and success measures
4. Recommended approach
5. Deliverables and exclusions
6. Timeline and responsibilities
7. Investment and terms
8. Assumptions and risks
9. Relevant proof
10. Clear next step

Write for the prospect's priorities, not our internal vocabulary. Keep claims specific and traceable to the notes.

15. Respond to a customer complaint and prevent a repeat

Use this to draft the customer reply and create a separate internal action plan.

Help resolve this customer complaint.

Customer message: [paste message]
Account context: [relevant facts only]
Verified issue: [what happened, if known]
Policy: [refund, service, warranty or escalation policy]
Actions we can offer: [approved options]
Tone: [brand voice]
Owner: [person or team responsible]

Do not admit legal liability, promise compensation or state a cause unless the supplied facts support it. Do not blame the customer or another team.

Return 2 separate sections:

CUSTOMER RESPONSE
- Acknowledge the specific problem
- State what is known without speculation
- Explain the next action, owner and timing
- Offer only approved remedies
- Keep it concise and human

INTERNAL FOLLOW-UP
- Issue summary
- Immediate containment
- Likely root causes to investigate
- Missing evidence
- Corrective actions with owners
- Whether similar customers may be affected
- Suggested CRM or support tags

Flag any sentence requiring legal, policy or management approval.

Operations and productivity prompts

16. Turn meeting notes into decisions and actions

This prompt separates what was decided from what was merely discussed. That distinction has saved more projects than another meeting ever will.

Turn these meeting notes or transcript into an accurate follow-up.

Meeting purpose: [purpose]
Participants: [names and roles]
Source: [paste notes or transcript]
Project context: [brief context]

Use only the source. Do not assign an owner, deadline or decision unless it was stated. Mark unclear speaker attribution as [UNCLEAR].

Return:
1. Executive summary in no more than 5 bullets
2. Decisions made, with decision owner where stated
3. Action items in a table: action, owner, due date, dependency and status
4. Open questions
5. Risks or blockers raised
6. Topics discussed but not decided
7. A draft follow-up message to participants

Quote the exact source line for any commitment that could be disputed. Put items with no owner or date in a separate "Needs assignment" section.

For repeatable extraction, try GPTBot’s note-taking assistant prompt.

17. Convert a process into an SOP

Use this when a recurring process lives in someone’s memory, a Slack thread and one ominously named spreadsheet.

Create a standard operating procedure from the process information below.

Process: [name and objective]
Current steps: [notes, transcript or screen recording summary]
Trigger: [what starts the process]
Inputs: [documents, data or approvals needed]
Systems: [tools used]
Roles: [people involved]
Output: [definition of complete]
Known exceptions: [unusual cases]
Policies or controls: [requirements]

If essential information is missing, ask up to 5 questions in one batch. Do not invent system behavior, permissions or approval rules.

Return:
1. Purpose and scope
2. Trigger, inputs and prerequisites
3. Roles and responsibilities
4. Numbered steps detailed enough for a trained new employee
5. Decision points and exception paths
6. Quality checks and evidence to retain
7. Common failures and recovery steps
8. Escalation path
9. Completion checklist
10. Version owner and review frequency

Mark every uncertain step as [CONFIRM]. Keep instructions direct and testable.

18. Build an executable project plan

This prompt converts an outcome into work packages without pretending every estimate is certain.

Create a project plan for this outcome.

Objective: [specific outcome]
Definition of done: [measurable completion criteria]
Deadline: [date]
Team: [people, roles and capacity]
Budget: [amount]
Constraints: [systems, approvals, policy or timing]
Known tasks: [list]
Dependencies: [list]
Risks: [known risks]

Return:
1. Scope and explicit exclusions
2. Work breakdown by phase
3. Milestones and acceptance criteria
4. Task table with owner, estimate, dependency and status
5. Critical path and bottlenecks
6. Risk register with probability, impact, mitigation and owner
7. Decision and approval points
8. Communication cadence
9. The first 7 days of work

Label estimates as estimates. Flag impossible sequencing, unowned work and missing capacity. If the deadline conflicts with the supplied effort, show the tradeoff instead of forcing the plan to fit.

19. Create a useful weekly status report

This turns project data into a briefing stakeholders can scan without digging through a task board.

Write a weekly project status report from the data below.

Project: [name and objective]
Audience: [stakeholders]
Reporting period: [dates]
Last report: [paste if available]
Completed work: [list]
Current work: [list]
Milestones: [status]
Metrics: [actual figures]
Risks and blockers: [list]
Decisions needed: [list]
Next period: [planned work]

Do not invent progress percentages, dates, owners or confidence. Compare with the previous report and call out meaningful changes.

Return:
1. Overall status: on track, at risk or off track, with one-sentence rationale
2. Outcomes completed, not activity for its own sake
3. Milestone table
4. Metric changes
5. Risks and blockers with owner and next action
6. Decisions needed, including decision maker and deadline
7. Priorities for next week

Keep it under 500 words. Put missing or contradictory information in a final "Data to confirm" section.

GPTBot’s project status report template prompt can help standardize the format.

20. Find high-value AI automation opportunities

This prompt looks for repetitive workflows worth improving. It also filters out automations that are technically cute and operationally useless.

Identify and rank AI automation opportunities in these workflows.

Business: [description]
Team: [roles and capacity]
Workflows: [list each workflow with steps]
Frequency and volume: [how often each runs]
Time spent: [estimates]
Tools and systems: [current stack]
Error or delay costs: [known impact]
Data sensitivity: [public, internal, confidential or regulated]
Approval requirements: [human checks]
Budget and technical capacity: [constraints]

For each workflow:
1. Identify repetitive, rules-based and judgment-heavy steps
2. Separate traditional automation from tasks that may benefit from AI
3. Estimate value using only supplied volume and time data
4. Assess implementation effort, data access, error impact and reversibility
5. Define required human review
6. Suggest a small pilot with success and stop criteria

Return a ranked table with value, effort, risk, confidence and recommended approach. Exclude workflows where a simple checklist, form or integration would solve the problem more reliably. Do not invent time savings.

Use GPTBot’s AI workflow opportunity finder for an interactive version.

Finance and people management prompts

21. Explain budget variances

Use this to turn actual budget data into questions and a management briefing. It is not a replacement for accounting review.

Analyze this budget-versus-actual report.

Period: [dates]
Budget data: [paste table or attach file]
Actual data: [paste table or attach file]
Prior period: [optional]
Materiality threshold: [amount or percentage]
Business context: [known events that affected results]
Audience: [leadership, department head or board]

Preserve the supplied figures exactly. Show formulas for calculations and verify totals. Do not invent causes for variances.

Return:
1. Executive summary
2. Material favorable and unfavorable variances
3. Table with budget, actual, absolute variance and percentage variance
4. Known explanations supported by context
5. Possible explanations clearly labeled as hypotheses
6. Questions for budget owners
7. Risks to the next period
8. Recommended actions with owners where supplied

Flag inconsistent categories, missing values and totals that do not reconcile. Keep factual analysis separate from recommendations.

A finance owner should verify every figure before the briefing leaves the draft folder.

22. Build cash-flow scenarios

This creates a transparent scenario model from your inputs. It should expose the assumptions, not bury them beneath a reassuring chart.

Build a cash-flow scenario analysis from the figures below.

Opening cash: [amount]
Forecast period: [months]
Expected cash inflows by month: [table]
Expected cash outflows by month: [table]
Payment timing assumptions: [terms and collection delays]
One-time costs: [list]
Debt or financing: [payments and available facilities]
Minimum cash buffer: [amount]
Known uncertainties: [late payments, seasonality, hiring or other factors]

Create base, downside and upside scenarios. Use only supplied figures and show every changed assumption.

Return:
1. Monthly cash table for each scenario
2. Lowest cash point and month
3. Cash runway where relevant
4. Months below the minimum buffer
5. The assumptions with the largest impact
6. Early warning indicators
7. Actions available before a shortfall
8. Missing data that would improve the model

Show formulas, reconcile opening and closing cash, and check the calculations. Do not provide tax, investment or legal advice.

23. Review pricing and design a validation plan

Use this when considering a price increase, packaging change or new pricing model.

Review these pricing options and design a validation plan.

Offer: [product or service]
Customer segments: [segments]
Current pricing and packaging: [details]
Unit costs and delivery capacity: [figures]
Sales data: [conversion, discounts, deal size, retention or other evidence]
Customer research: [quotes, objections and willingness-to-pay evidence]
Competitor pricing: [verified data and dates]
Options under consideration: [list]
Commercial goal: [margin, adoption, expansion or simplicity]
Constraints: [contracts, channel, positioning or systems]

Do not invent willingness to pay or recommend a price based only on competitor averages.

Return:
1. Diagnosis of the current pricing
2. Segment-level value drivers
3. Comparison of each option across customer value, margin, complexity and risk
4. Unit economics using supplied figures
5. Customers likely to benefit or lose under each option
6. A validation plan using interviews, sales tests or controlled experiments
7. Success, guardrail and stop metrics
8. Rollout and communication risks
9. Recommendation and confidence level

List the assumptions that must be true for the recommendation to work.

24. Create a hiring scorecard and interview kit

This keeps interviews focused on evidence of future job performance instead of chemistry and confident storytelling.

Create a structured hiring scorecard and interview kit for this role.

Role: [title]
Business context: [company, team and stage]
Outcomes required in the first 12 months: [list]
Responsibilities: [list]
Must-have skills: [list]
Trainable skills: [list]
Constraints: [location, schedule, salary range or legal requirements]
Interview stages: [list]

Base the scorecard on job-relevant outcomes and observable behavior. Remove criteria that are vague, duplicative or unrelated to performance.

Return:
1. Role mission in 2 sentences
2. 5 to 8 measurable outcomes
3. Competencies with behavioral anchors for scores 1, 3 and 5
4. Interview questions mapped to each criterion
5. Follow-up probes that request specific examples
6. A practical work-sample exercise with scoring rubric
7. A consistent interviewer score sheet
8. Red flags that require evidence, not intuition
9. Questions to check candidate expectations and role fit

Do not infer protected characteristics or recommend discriminatory criteria. Flag any requirement that may need HR or legal review.

GPTBot also offers an Act as a Recruiter prompt.

25. Prepare a balanced performance review

Use this to organize evidence and plan a clear conversation. Do not let ChatGPT decide a rating, promotion or disciplinary outcome.

Help me prepare a fair performance review from the evidence below.

Employee role and level: [details]
Review period: [dates]
Agreed goals: [goals]
Documented outcomes: [results and metrics]
Feedback received: [specific feedback with context]
Examples of strengths: [examples]
Examples of missed expectations: [examples]
Role expectations: [approved framework]
Employee self-review: [optional]
Company review format: [sections]

Use only documented evidence. Do not infer intent, personality or protected characteristics. Separate outcomes from circumstances outside the employee's control.

Return:
1. Evidence mapped to each agreed goal
2. Strengths with specific examples and business impact
3. Gaps with specific examples and expected standard
4. Contradictory or incomplete evidence
5. A balanced draft in the company format
6. Questions to understand the employee's perspective
7. A development plan with measurable next steps and support required
8. A conversation outline for the manager
9. Statements requiring HR review

Do not recommend a compensation, promotion or termination decision. Flag where policy or human judgment is required.

A quick review checklist before using ChatGPT output

Before sending, publishing or acting on the response, check:

• Are the names, dates, figures, quotations and links correct?
• Did ChatGPT use supplied evidence or quietly fill gaps?
• Are assumptions labeled?
• Does the output make unsupported promises or claims?
• Is confidential or personal information handled according to policy?
• Does a qualified person need to review legal, financial, HR, security or technical content?
• Is the proposed next action realistic for the people who must execute it?

The prompt is the brief. The output is still a draft.

Frequently asked questions

What is the best ChatGPT prompt for business?

The best prompt is tied to a recurring task with a clear input and review standard. For many teams, meeting actions, customer feedback analysis, sales follow-ups and SOP drafting are good starting points because the source material already exists and a person can easily judge the result.

How do I stop ChatGPT from giving generic business advice?

Provide evidence from your business: customer language, actual constraints, current performance and examples of approved work. Ask ChatGPT to separate facts from assumptions and identify missing information instead of filling every gap.

Can ChatGPT make business decisions for me?

It can organize evidence, compare options, challenge assumptions and draft a recommendation. The decision still belongs to the person accountable for its consequences. This matters most when the choice involves employees, customers, contracts, safety or significant money.

Can I paste confidential business information into ChatGPT?

Follow your organization’s policy and the controls of the plan you use. OpenAI documents separate controls for business products and consumer ChatGPT accounts. Remove unnecessary personal or confidential data, use approved accounts and tools, and involve security or legal owners when the data is sensitive or regulated.

Do these prompts work with Claude, Gemini, Grok and other AI assistants?

Yes. The same prompts generally work with Claude, Gemini, Grok, Microsoft Copilot and other capable AI assistants because they use model-agnostic instructions: define the task, provide context and evidence, specify constraints, request a format and review the result. Adjust them when an assistant has different tools, file limits, web access or connected data sources.

Start with one recurring task

Pick the task your team repeats every week, fill in one prompt with real context and run it. Review the result, tighten the weak instruction and save the version that works.

Browse the full GPTBot ChatGPT prompt library when you need a more specialized template.