Every week I get about 50 questions from readers. Most fall into the same categories: “How do I use AI for X?” or “Which tool is best for Y?” or “Can you share the prompt you use for Z?”
This week I’m answering seven of the most common questions. Not with theory. With actual prompts I’ve tested at least 15 times each, along with real outputs so you can see what good looks like.
Copy these. Use them. Adjust them for your situation. That’s the point.
Question 1: Social Media Content That Doesn’t Sound Like AI
“Jordan, how do you create social posts that don’t scream ‘I used ChatGPT’? Every time I try, it comes out generic and boring. What’s your process?”
- Marcus T., B2B SaaS
The Problem: AI defaults to safe, corporate-speak nonsense. “Excited to announce...” and “Thrilled to share...” and other phrases no human actually uses in conversation.
The Solution: Give it constraints that force personality. Here’s my exact prompt for LinkedIn posts:
LinkedIn Post Prompt (ChatGPT or Claude):
“Write a LinkedIn post about [TOPIC] that follows these rules:
1. Start with a one-sentence hook that creates curiosity or tension. No questions. No ‘imagine this’ setups.
2. Use short sentences. Nothing over 15 words. Write like you talk.
3. Include one specific number, stat, or timeframe to add credibility.
4. End with a clear takeaway, not a question asking for engagement.
5. Never use these phrases: ‘excited to announce’, ‘thrilled to share’, ‘diving deep’, ‘game changer’, ‘passionate about’, ‘circling back’
Topic: [INSERT YOUR TOPIC]
Tone: Direct, slightly provocative, like you’re talking to a peer over coffee
Length: 150-200 words max”
Example Output (Topic: Why most AI adoption fails):
Most companies are wasting six figures on AI tools they’ll abandon in 90 days.
Here’s why. They buy the software before defining the problem. They train people on features instead of outcomes. They expect magic without changing workflows.
I watched a client spend $180K on an AI sales platform last year. Three months later, adoption rate was 12%. The tool wasn’t bad. Their approach was.
What actually works: Start small. Pick one painful manual process. Automate that. Measure the result. Then scale.
We did this with lead enrichment. Saved 18 hours weekly. Cost $178 monthly. ROI was obvious, so the team actually used it.
Stop buying solutions to problems you haven’t defined. Start with the problem. The tool is step five, not step one.
Why this works: Specific constraints force the AI away from generic patterns. The banned phrases list is critical. Update it with whatever phrases annoy you in AI content.
Best model for this: ChatGPT for casual tone, Claude if you want more structure.
Question 2: Contract Review Without Paying Your Lawyer
“I get sent contracts and NDAs constantly. Can AI actually review these reliably, or am I risking something by not using a lawyer every time?”
- Sarah K., Consultant
The Reality: AI can’t replace lawyers for high-stakes contracts. But for standard NDAs, service agreements, and vendor contracts under $25K, it’s shockingly good at flagging red flags.
The Prompt (Claude works best for this):
Contract Review Prompt:
“You are a contract attorney reviewing this agreement on behalf of my company. Analyze this contract and tell me:
1. Red flags: Any clauses that heavily favor the other party or expose me to unusual risk
2. Standard vs. non-standard: Which terms are typical for this type of agreement and which are unusual
3. Missing protections: What a contract like this should include that’s not here
4. Key dates and obligations: Deadlines, renewal terms, termination clauses
5. Financial exposure: All ways this contract could cost me money beyond the stated fees
Format your response as: Executive Summary (2-3 sentences), then the 5 sections above.
[PASTE CONTRACT TEXT]”
What I’ve caught using this:
• Auto-renewal clause buried in section 12 of a vendor agreement (would’ve cost me $18K/year I didn’t budget for)
• Unlimited liability clause in a standard NDA (my lawyer laughed when I sent it to him)
• Non-compete that would’ve prevented me from working with anyone in my industry for 2 years
• Payment terms that required net-15 but gave them net-90 (spotted the asymmetry immediately)
Important disclaimer: I still send anything over $25K or anything involving IP/equity to my actual lawyer. AI catches 80% of issues in standard contracts. For the other 20%, you need human expertise.
Cost saved: My lawyer charges $400/hour. A basic contract review is $800 minimum. I’ve used this prompt 23 times. Sent only 4 to my lawyer for final review. Saved roughly $15,200 in legal fees.
Question 3: Sales Scripts That Convert
“Can AI write sales scripts for cold calls or discovery calls? I need something that sounds natural, not robotic.”
- Devon R., Sales Leader
The Approach: AI can’t write a perfect sales script from scratch. But it’s great at creating frameworks and handling objections if you give it context.
The Prompt (ChatGPT):
Discovery Call Script Prompt:
“Create a discovery call script for selling [YOUR PRODUCT/SERVICE] to [TARGET CUSTOMER]. Structure it as:
1. Opening (30 seconds): Build rapport without small talk, transition to business
2. Qualification Questions (5 questions): Uncover budget, timeline, pain points, decision process
3. Value Framing (2 minutes): Connect their pain to your solution, use a brief case study
4. Objection Handling: Responses to the 3 most common objections: [LIST YOUR COMMON OBJECTIONS]
5. Close: Next step that moves the deal forward
Make it conversational. Use transitions that feel natural. Don’t make it sound like a script someone is reading.
Product: [DESCRIBE]
Target Customer: [DESCRIBE]
Typical Pain Points: [LIST]
Common Objections: [LIST]”
Example Output (B2B SaaS sales):
Opening:
“Hey [Name], thanks for taking the time. I know you’re busy, so I’ll keep this focused. I wanted to understand more about how your team currently handles [PAIN POINT] and see if what we’ve built might be a fit. Sound good?”
Qualification Questions:
1. “Walk me through your current process for [TASK]. What’s working and what’s frustrating?”
2. “How much time does your team spend on this weekly?”
3. “Have you looked at other solutions? What stopped you from pulling the trigger?”
4. “If we could solve this, what would success look like in 90 days?”
5. “Who else needs to be involved in evaluating a solution like this?”
Pro tip: Use the output as a starting point. Record yourself reading it. Listen back. Fix anything that sounds awkward. Then practice it 10 times until it’s natural.
Question 4: Data Analysis for Non-Technical People
“I have spreadsheets full of customer data but no clue how to analyze it. Can AI help me find patterns without needing to learn Excel formulas?”
- Jennifer L., Product Manager
The Answer: Yes, and it’s one of AI’s best use cases. You don’t need technical skills. You just need to ask the right questions.
The Prompt (Claude or ChatGPT with data analysis):
Data Analysis Prompt:
“I have a dataset with [DESCRIBE COLUMNS]. Analyze this data and tell me:
1. Top 3 patterns or trends
2. Any surprising outliers or anomalies
3. Segments or clusters in the data (group similar customers/responses)
4. Actionable recommendations based on what you found
Present findings in plain English. No technical jargon. Explain your reasoning.
[PASTE DATA or UPLOAD FILE]”
Real example I ran last month:
Dataset: 312 customer survey responses about our product (ratings 1 to 10 on various features, plus open text feedback)
What Claude found:
• Customers rating onboarding below 7 had 68% higher churn in first 90 days
• The #1 feature request (integrations) was mentioned by 41% of respondents, but only 12% used our existing integrations
• Customers in healthcare vertical rated us 2.3 points lower on average, citing compliance concerns we hadn’t addressed
• NPS promoters used an average of 4.2 features, detractors used 1.8 features
Actions we took: Rebuilt onboarding, created healthcare-specific compliance docs, focused on driving multi-feature adoption. NPS jumped 11 points in 8 weeks.
Cost to get those insights from a data analyst: $3K to $5K. Cost using AI: $0 (already had ChatGPT subscription).
Question 5: Email Sequences That Don’t Get Ignored
“I need to build a 5-email nurture sequence for leads who downloaded our guide. How do I make each email valuable instead of just ‘following up’?”
- Alex P., Marketing Director
The Framework: Each email needs a single clear purpose and a reason to exist. AI can help structure this if you give it the right constraints.
Email Sequence Prompt:
“Create a 5-email nurture sequence for people who downloaded [LEAD MAGNET]. Each email should:
- Have a specific purpose (education, objection handling, social proof, urgency, or conversion)
- Provide new value, not just rehash the previous email
- Be 150-200 words max
- Include a single clear CTA
Sequence timing: Day 0, Day 3, Day 7, Day 12, Day 18
Lead magnet: [DESCRIBE]
Product being sold: [DESCRIBE]
Target audience: [DESCRIBE]
Main objections: [LIST]”
Output structure you’ll get:
Email 1 (Day 0): Deliver the promised resource, set expectations for the sequence
Email 2 (Day 3): Share a case study or success story related to the guide’s topic
Email 3 (Day 7): Address the #1 objection with data or proof
Email 4 (Day 12): Introduce your product as the natural next step
Email 5 (Day 18): Create soft urgency (bonus, limited spots, ending soon)
Pro tip: Don’t use the AI output verbatim. Use it as scaffolding. Add your voice, your examples, your personality. The structure is solid, but it needs your human touch to convert.
Question 6: Meeting Prep That Actually Helps
“I have back-to-back client meetings and never enough time to prep. Can AI help me get ready faster?”
- Tom W., Agency Owner
The System: Feed AI your calendar, previous meeting notes, and client context. Get a prep brief in 90 seconds.
Meeting Prep Prompt:
“I have a meeting with [CLIENT NAME] in [TIME]. Create a meeting prep brief that includes:
1. Meeting objective (what should I accomplish?)
2. Key discussion points (3-5 topics to cover)
3. Questions to ask them
4. Potential objections or concerns they might raise
5. Desired outcome and next steps
Context about this client:
- Previous interactions: [SUMMARIZE or PASTE NOTES]
- Their current situation: [DESCRIBE]
- What they’ve bought from us: [LIST]
- Known pain points: [LIST]
Keep it under 300 words. Focus on what I need to know to make this meeting productive.”
When I use this: Every client meeting. Takes me 2 minutes to fill in the context, AI gives me a brief in 30 seconds. I read it while walking to the meeting or waiting for them to join the Zoom.
Result: I show up prepared. I reference things they mentioned two months ago. I ask questions that move deals forward instead of spinning wheels. Clients notice.
Question 7: Repurposing Content Without Sounding Repetitive
“I write one long-form piece per week but need to turn it into social posts, email, and video script. How do I repurpose without sounding like I’m saying the same thing everywhere?”
- Rachel M., Content Creator
The Strategy: Different platforms need different angles. AI can extract multiple perspectives from one piece.
Content Repurposing Prompt:
“I wrote this article: [PASTE ARTICLE or SUMMARY]. Repurpose it into:
1. LinkedIn post (200 words, focus on the main insight)
2. Twitter thread (5 tweets, focus on the contrarian take)
3. Email newsletter intro (150 words, focus on why readers should care)
4. 60-second video script (focus on the problem and solution)
5. Instagram caption (100 words, focus on the emotional hook)
Make each one feel native to the platform. Don’t just shrink the article. Find different angles.”
Pro move: Run this through Galaxy.ai and compare outputs from different models. ChatGPT usually nails the casual platforms (Twitter, Instagram). Claude does better with professional stuff (LinkedIn, email).
I’ve been using this workflow for 4 months. One article becomes 15+ pieces of content. Total time: 20 minutes including review and editing.
Tools to Make This Easier:
All these prompts work better when your workflow is dialed in. Here’s my current stack:
For social scheduling: Buffer. I generate all my social posts on Monday, schedule them for the week, forget about it.
For newsletters: Beehiiv. If you’re still using Mailchimp, stop. Beehiiv has better deliverability, cleaner interface, and actual growth tools.
For multi-model prompting: Galaxy.ai (covered this Tuesday, still the best way to compare outputs across models)
The Quality Benchmark:
How do you know if AI output is good enough to use? Here’s my filter:
Good output:
• Sounds like something a human would actually say
• Has specific details, not vague generalities
• Makes a clear point without hedging
• Flows naturally when read aloud
• Doesn’t include phrases you’d never use
Bad output (regenerate or heavily edit):
• Overly formal or corporate tone
• Lists obvious points everyone already knows
• Uses clichés or buzzwords
• Feels like it was written by committee
• You can tell it’s AI within the first sentence
The 10-second test: Read it out loud. If you’d say it to a colleague, it’s good. If it sounds like a press release, regenerate.
Send Me Your Questions:
I’m doing this Q&A format monthly. Send your questions to [email protected] with subject line “Q&A Question”. I’ll test them, document the results, and share what works.
Focus on specific use cases. “How do I use AI for marketing?” is too broad. “How do I use AI to write better cold email subject lines?” is perfect.
Jordan Hale
The AI Newsroom
P.S. All these prompts (plus 53 more) are organized by use case in the AI Workflow Blueprint ($47). You get the prompts, the workflows they plug into, and video walkthroughs showing exactly how to implement them. Get it today by commenting or replying “BLUEPRINT”