How to Craft Prompts for AI Models

🤖 Hey fellow medical writers! After months of experimenting with AI tools, I wanted to share some game-changing prompting tips that have revolutionized my workflow. Let's get real - crafting the perfect prompt is an art, and in medical writing, the stakes are high. Here's what I've learned: 1. Get super specific! 📍 Instead of "write about diabetes," try "explain the latest Type 2 diabetes treatment guidelines (uploaded document X) for newly diagnosed patients." The more precise you are, the better your results. 2. Context is your best friend 🤝 I always include my target audience, purpose, and any specific requirements. For example: "This is for healthcare providers at a community clinic who need quick, actionable information during patient visits." 3. Break it down to build it up 🏗️ Tackle complex projects in bite-sized pieces. If I am writing a needs assessment - instead of asking for the whole thing at once, I brake it into smaller tasks: research gathering, summary, and outline development. Game changer! 4. Embrace the back-and-forth 🔄 Perfect prompts rarely happen on the first try. I treat it like a conversation, refining my requests based on what the AI gives me. It's all about that iterative process! 5. Keep it human 🧠 Remember - AI is a tool, not a replacement. I always fact-check, review for accuracy, and ensure the content aligns with ethical guidelines. Trust but verify! Pro tip: I keep a "prompt library" of my most successful prompts. It's like having a secret weapon in my medical writing toolkit! What are your favorite prompting strategies? Drop them in the comments below! 👇 #MedicalWriting #AI #HealthcareCommunication #WritingTips #ProfessionalDevelopment

🧠 Designing AI That Thinks: Mastering Agentic Prompting for Smarter Results Have you ever used an LLM and felt it gave up too soon? Or worse, guessed its way through a task? Yeah, I've been there. Most of the time, the prompt is the problem. To get AI that acts more like a helpful agent and less like a chatbot on autopilot, you need to prompt it like one. Here are the three key components of an effective 🔁 Persistence: Ensure the model understands it's in a multi-turn interaction and shouldn't yield control prematurely. 🧾 Example: "You are an agent; please continue working until the user's query is resolved. Only terminate your turn when you are certain the problem is solved." 🧰 Tool Usage: Encourage the model to use available tools, especially when uncertain, instead of guessing. 🧾 Example:" If you're unsure about file content or codebase structure related to the user's request, use your tools to read files and gather the necessary information. Do not guess or fabricate answers." 🧠 Planning: Prompt it to plan before actions and reflect afterward. Prevent reactive tool calls with no strategy. 🧾 Example: "You must plan extensively before each function call and reflect on the outcomes of previous calls. Avoid completing the task solely through a sequence of function calls, as this can hinder insightful problem-solving." 💡 I've used this format in AI-powered research and decision-support tools and saw a clear boost in response quality and reliability. 👉 Takeaway: Agentic prompting turns a passive assistant into an active problem solver. The difference is in the details. Are you using these techniques in your prompts? I would love to hear what's working for you; leave a comment, or let's connect! #PromptEngineering #AgenticPrompting #LLM #AIWorkflow

Over the last month, I've been talking to founders building AI companies, and I realized something embarrassing: I've been doing AI prompting wrong. Like many others, I used to: ↳ Dump everything into a single prompt ↳ Add some basic instructions ↳ Hope for the best Then I discovered what top AI builders do differently: They master the art of "context sculpting." (Thanks David Wilson/Hunch for sharing this term with me) What is context sculpting? ↳ Strategically structuring and organizing context for AI models ↳ Using the right formatting to separate different types of information ↳ Carefully selecting and curating examples Here's what I've learned from them: 1/ Structure beats chaos XML tags consistently outperform markdown or plain text. For example: <context>Your base information</context> <examples>Your carefully chosen examples</examples> <task>Your specific instructions</task> 2/ Example selection is critical More examples ≠ better results Each example you include: ↳ Increases token costs ↳ Can "poison" your prompt with unwanted patterns ↳ Must be carefully chosen to represent desired outputs 3/ The power of iteration Start minimal. Test outputs. Add or remove context strategically. Find the sweet spot between too little and too much. So next time you're prompting your favorite AI model: 1. Structure your context with XML tags 2. Be selective with examples 3. Iterate on context, not just instructions 4. Test different organization patterns I'll be surprised if the outputs you get back are not substantially better. What prompting techniques have worked for you?

You’re doing it. I’m doing it. Your friends are doing it. Even the leaders who deny it are doing it. Everyone’s experimenting with AI. But I keep hearing the same complaint: “It’s not as game-changing as I thought.” If AI is so powerful, why isn’t it doing more of your work? The #1 obstacle keeping you and your team from getting more out of AI? You're not bossing it around enough. AI doesn’t get tired and it doesn't push back. It doesn’t give you a side-eye when at 11:45 pm you demand seven rewrite options to compare while snacking in your bathrobe. Yet most people give it maybe one round of feedback—then complain it’s “meh.” The best AI users? They iterate. They refine. They make AI work for them. Here’s how: 1. Tweak AI's basic setting so it sounds like you AI-generated text can feel robotic or too formal. Fix that by teaching it your style from the start. Prompt: “Analyze the writing style below—tone, sentence structure, and word choice—and use it for all future responses.” (Paste a few of your own posts or emails.) Then, take the response and add it to Settings → Personalization → Custom Instructions. 2. Strip Out the Jargon Don’t let AI spew corporate-speak. Prompt: “Rewrite this so a smart high schooler could understand it—no buzzwords, no filler, just clear, compelling language.” or “Use human, ultra-clear language that’s straightforward and passes an AI detection test.” 3. Give It a Solid Outline AI thrives on structure. Instead of “Write me a whitepaper,” start with bullet points or a rough outline. Prompt: “Here’s my outline. Turn it into a first draft with strong examples, a compelling narrative, and clear takeaways.” Even better? Record yourself explaining your idea; paste the transcript so AI can capture your authentic voice. 4. Be Brutally Honest If the output feels off, don’t sugarcoat it. Prompt: “You’re too cheesy. Make this sound like a Fortune 500 executive wrote it.” or “Identify all weak, repetitive, or unclear text in this post and suggest stronger alternatives.” 5. Give it a tough crowd Polished isn’t enough—sometimes you need pushback. Prompt: “Pretend you’re a skeptical CFO who thinks this idea is a waste of money. Rewrite it to persuade them.” or “Act as a no-nonsense VC who doesn’t buy this pitch. Ask 5 hard questions that make me rethink my strategy.” 6. Flip the Script—AI Interviews You Sometimes the best answers come from sharper questions. Prompt: “You’re a seasoned journalist interviewing me on this topic. Ask thoughtful follow-ups to surface my best thinking.” This back-and-forth helps refine your ideas before you even start writing. The Bottom Line: AI isn’t the bottleneck—we are. If you don’t push it, you’ll keep getting mediocrity. But if you treat AI like a tireless assistant that thrives on feedback? You’ll unlock content and insights that truly move the needle. Once you work this way, there’s no going back.

Remember Amelia Bedelia? The children's book character who, when asked to "draw the drapes," literally sketched a picture of curtains instead of closing them? That's exactly how AI sourcing tools work— and why most recruiters get terrible results from them. Like Amelia, AI takes your instructions literally. It can't always: - Read between the lines - Make judgment calls - Infer meaning from context - Understand subjective qualities This is why vague prompts like "find experienced backend developers" or "source high-performing sales leaders" return garbage candidates. So here’s how to write prompts that actually work: BAD PROMPT: "Find backend developers with: - Strong leadership skills - Extensive cloud experience - Proven track record of success" GOOD PROMPT: "Find backend developers who: - Led engineering teams of 5+ people for at least 2 years - Deployed and managed cloud infrastructure on AWS/Azure with budgets over $100K - Scaled systems handling 1M+ daily active users - Reduced infrastructure costs by at least 20% through optimization - Contributed to 3+ open source projects in the last 18 months" The difference? The second prompt gives AI concrete, measurable criteria it can evaluate from candidate profiles. 3 rules I follow for every AI sourcing prompt: 1. Replace subjective qualities with objective metrics "Leadership skills" → "Managed X people for Y years with Z outcomes" 2. Clarify work history "5 years of experience" → "X+ years as a financial analyst at a multi-national corporation" 3. Quantify impact wherever possible "Experienced in sales" → "X+ years of experience in SaaS sales with a consistent track record of exceeding quotas by YY%." PS: If you want to dive deeper into this, the Gem team dropped our complete playbook on mastering AI sourcing. Read it now here: https://lnkd.in/gSXeHCfV

Tara Behrend and I have just published these fully open-access guidelines, plus accompanying code for Qualtrics, to use LLMs/AI to create custom content for surveys and experiments, quantitative or qualitative! The code for Qualtrics is as close to plug-and-play as we could make it, only requiring one copy-paste followed by changing a few settings at the top of the code block. It enables researchers to easily: 1) Create unique AI-generated content per participant (Case 2) 2) Engage participants in an LLM-based conversation with a researcher-designed system prompt (Case 4) 3) Experimentally assign participants to different LLM configurations (Case 5) My hope is that this tool increases access to LLMs for social scientists of all backgrounds. All you need is a Qualtrics account (provided for free by many universities) and a OpenAI API key. Research studies with a few hundred participants will generally cost less than $5 in API credits from OpenAI. Beyond the software itself, we developed a framework for the general use of LLMs to create content for research participants to experience/react to: Case 1) LLM as Research Assistant Case 2) LLM as Adaptive Content Provider Case 3) LLM as External Resource Case 4) LLM as Conversation Partner Case 5) LLM as Research Confederate Across cases, we provided detailed instructions on how to effectively engineer an LLM for research, including an iterative design thinking framework for prompt engineering and foundation model specification, as well as recommendations for a comprehensive audit before launch. We also present a nine-dimensional model of prompt design alongside recommendations for how to create effective prompts for research! I hope you find it useful, and I'm happy to help troubleshoot as you explore it! https://lnkd.in/gwtfH-HG

I’ve been getting a ton of DMs and comments on my last post asking about the AI prompt I used to build an executive hiring assessment and interview plan. Not surprising, executive search is still broken in a lot of places. Too many interviews rely on instinct, vague questions, or outdated playbooks. So instead of sending my prompt to everyone individually, I figured I’d walk through how to build one yourself. Most prompts fall flat because they’re too shallow. A strong AI prompt for executive hiring should be treated like an API request: structured, layered, and loaded with context. Here’s how to build one that actually works: 1. Start with Context Give the AI business and hiring context, not just the role title. “You’re designing an interview plan for a Series B SaaS company hiring a VP of Marketing to scale pipeline efficiency and launch EMEA.” 2. Provide Source Material Feed the AI what you already have. Include: - Job description - Job ad - Talent Ladder - Company values - Success profiles or leveling framework - Team charter or hiring manager notes - Previous document structures that have worked in the past - Anything that hasn’t worked in the past Label each section clearly. The more signal you give, the more signal you get. 3. Define the Output Format Tell the AI what you want. Example: “Build a 5-stage interview loop. For each stage, include: - Interview title and purpose - Core competencies - Suggested questions (behavioral + scenario-based) - Evaluation criteria - Interviewer profile - Timebox in minutes” Structure = clarity. 4. Add Constraints + Nuance Make the AI smarter by telling it what to avoid and what to emphasize. “Avoid generic leadership questions. Emphasize org design, systems thinking, and cross-functional collaboration. Deprioritize ‘culture fit’ in favor of ‘culture add.’” 5. Layer in Iteration AI isn’t one-and-done. After the first output, ask: “Now revise this plan to reflect our values more clearly.” “Make this inclusive of candidates from non-traditional backgrounds.” “Sharpen the rubric for strategic thinking.” Bonus tip: Once you’ve dialed in your prompt structure, try formatting it in YAML. It makes the logic cleaner, easier to reuse, and more adaptable for tools like Gemini or ChatGPT. If you build your prompt with this kind of structure, you’ll stop getting interview plans that look smart and start getting ones that actually work. And yes, AI can do incredible things 💜 but only if you feed it the right ingredients. Have questions or want to troubleshoot your own prompt design? Happy to trade notes in the comments. #executivesearch #promptengineering #AI #talentacquisition #recruiting

AI isn't ruining your resume. 🤖 Your prompts are. Here's the truth: Every recruiter knows generic AI-written resumes. They're flooding inboxes daily. But AI used RIGHT? That's your secret weapon. The difference is in HOW you prompt. 5 techniques to keep YOUR voice while leveraging AI: 1️⃣ The 'Voice Mirror' Method DON'T: 'Write a resume bullet for project management' DO: 'Here's how I naturally describe my work: [insert your casual explanation]. Now help me transform this into a metric-driven resume bullet while keeping my conversational tone.' 2️⃣ The 'Story First' Approach   Feed AI your RAW story first: 'I saved my team from a disaster when our vendor ghosted us before launch. I found 3 new vendors in 48 hours...' THEN ask for resume bullets. 3️⃣ The 'Mad Libs' Technique Prompt: 'I need to show I [specific action] that resulted in [specific outcome] for [specific stakeholder]. Give me 5 ways to say this authentically.' 4️⃣ The 'Peer Review' Hack 'Pretend you're my colleague describing my work to our boss. How would you highlight this achievement: [your achievement]?' 5️⃣ The 'Iteration Loop' • Start with AI draft • Add YOUR specific details • Ask AI to enhance while maintaining your additions • Repeat until it sounds like YOU Pro Framework for ANY resume prompt: Context + Specifics + Tone + Constraints = Authentic AI Output Example: 'I'm a marketing manager (context) who increased email conversions by 47% using A/B testing (specifics). Write this in a confident but not boastful tone (tone) in under 50 words (constraints).' The goal isn't to hide AI use. It's to use AI like spell-check on steroids. 📝 Your experiences are unique. Your voice should be too. AI is your writing assistant, not your ghostwriter. Start with YOUR story. Let AI help you tell it better. Perfect your AI prompts with Teal's Resume Builder guidance → https://lnkd.in/gJSNk4FN ♻️ Reshare to help someone write authentically. 🔔 Follow me for more job search & resume tips. #ResumeTips #AITools #JobSearch #CareerAdvice #ResumeWriting

Sometimes, prompting can feel like another job. You tried AI once, got absolute garbage back, and thought, "This is over-hyped nonsense." 𝗥𝗲𝗮𝗹𝗶𝘁𝘆 𝗰𝗵𝗲𝗰𝗸: It wasn't the AI...It was your prompt. 𝗛𝗲𝗿𝗲'𝘀 𝘄𝗵𝗮𝘁 𝗜'𝘃𝗲 𝗹𝗲𝗮𝗿𝗻𝗲𝗱: 𝗧𝗵𝗲 𝗽𝗿𝗼𝗺𝗽𝘁 𝘂𝘀𝘂𝗮𝗹𝗹𝘆 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 𝗺𝗼𝗿𝗲 𝘁𝗵𝗮𝗻 𝘁𝗵𝗲 𝗺𝗼𝗱𝗲𝗹. You know how to delegate to people, right? This is just learning how to delegate to software. Let me fix this for you. 𝗧𝗵𝗲 𝗗.𝗘.𝗙.𝗜.𝗡.𝗘. 𝗣𝗿𝗼𝗺𝗽𝘁 𝗦𝘆𝘀𝘁𝗲𝗺 I use this framework for every single prompt I write. It works across ChatGPT, Claude, Gemini - all of them. 𝗗 – Describe the Role   Don't say "help me." Say, "You are a senior consultant with expertise in nonprofit leadership." 𝗘 – Explain the End Goal   "Create a 1-page board summary," not "write something about our changes." 𝗙 – Focus the Context   Give background, but don't data dump. Relevant info only. 𝗜 – Include Inputs   Paste your notes, examples, and raw materials. Don't make AI guess what you want. 𝗡 – Narrow the Style & Format   "Calm, strategic voice. Under 300 words. Bullet points." Be specific. 𝗘 – Examples of Excellence   Show it what good looks like. "Here's last time's version - match that clarity." 𝗠𝘆 3-𝗦𝘁𝗲𝗽 𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄 (𝗧𝗵𝗮𝘁 𝗔𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗪𝗼𝗿𝗸𝘀) 𝗦𝘁𝗲𝗽 1 – 𝗣𝗶𝗰𝗸 𝗢𝗻𝗲 𝗥𝗲𝗮𝗹 𝗧𝗮𝘀𝗸 Weekly email, meeting prep, LinkedIn update - whatever's on your friction list. Run it through D.E.F.I.N.E. once. 𝗦𝘁𝗲𝗽 2 – 𝗨𝘀𝗲 𝗥.𝗜.𝗩. 𝗟𝗼𝗼𝗽 When it's 70% there (not perfect): → 𝗥𝗲𝘃𝗶𝗲𝘄: What's missing? → 𝗜𝘁𝗲𝗿𝗮𝘁𝗲: Change ONE thing → 𝗩𝗮𝗹𝗶𝗱𝗮𝘁𝗲: Run again, compare 2-3 rounds max. Done. 𝗦𝘁𝗲𝗽 3 – 60-𝗦𝗲𝗰𝗼𝗻𝗱 𝗦𝗮𝗳𝗲𝘁𝘆 𝗖𝗵𝗲𝗰𝗸 → Facts correct? → Sounds like you? → No confidential data? 𝗠𝗼𝘀𝘁 𝗽𝗲𝗼𝗽𝗹𝗲 𝗽𝗿𝗼𝗺𝗽𝘁 𝗹𝗶𝗸𝗲 𝘁𝗵𝗲𝘆 𝘁𝗮𝗹𝗸 - 𝘂𝗻𝗰𝗹𝗲𝗮𝗿, 𝘂𝗻𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲𝗱, 𝘁𝗼𝗼 𝗳𝗮𝘀𝘁. But you're building your future. Treat AI like your personal consultant: give it smart, focused instructions and watch the magic happen. Stop "playing with AI" and start succeeding with it. What's the one task taking up your time that you'd love to hand off to AI? 𝗪𝗮𝗻𝘁 𝗺𝗼𝗿𝗲 𝗔𝗜 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝘁𝗵𝗮𝘁 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘄𝗼𝗿𝗸? 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝘁𝗼 𝗺𝘆 𝘄𝗲𝗲𝗸𝗹𝘆 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 for practical AI frameworks (no fluff, I promise) 𝗙𝗼𝗹𝗹𝗼𝘄 𝗺𝗲 for daily insights on AI, productivity, and building smarter workflows 𝗗𝗠 𝗺𝗲 if you're building AI systems for your team - I'd love to hear what you're working on

𝐁𝐞𝐭𝐭𝐞𝐫 𝐀𝐈 𝐒𝐭𝐚𝐫𝐭𝐬 𝐰𝐢𝐭𝐡 𝐁𝐞𝐭𝐭𝐞𝐫 𝐏𝐫𝐨𝐦𝐩𝐭𝐬 Most people think prompt engineering is about clever wording. In reality? It’s strategic thinking in disguise. I use this simple framework “PROMPT” to write questions that actually work with AI: 𝗣𝘂𝗿𝗽𝗼𝘀𝗲𝗳𝘂𝗹 | 𝗥𝗲𝗹𝗲𝘃𝗮𝗻𝘁 | 𝗢𝗽𝗲𝗻-𝗲𝗻𝗱𝗲𝗱 | 𝗠𝗲𝗮𝗻𝗶𝗻𝗴𝗳𝘂𝗹 | 𝗣𝗿𝗲𝗰𝗶𝘀𝗲 | 𝗧𝗲𝘀𝘁 & 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗲 Here’s the breakdown: ➤ 𝗣𝘂𝗿𝗽𝗼𝘀𝗲𝗳𝘂𝗹 Start with why. Are you trying to brainstorm, summarize, or analyze?  Set a clear intention before you type a single word. ➤ 𝗥𝗲𝗹𝗲𝘃𝗮𝗻𝘁 Give the AI context. Keep it tightly focused. One misplaced detail can send it spinning into the wrong zone. ➤ 𝗢𝗽𝗲𝗻-𝗲𝗻𝗱𝗲𝗱 Ask questions that begin with how, why, or tell me about.  Avoid yes/no dead ends. ➤ 𝗠𝗲𝗮𝗻𝗶𝗻𝗴𝗳𝘂𝗹 Make it count. Don’t ask generic filler questions, craft prompts that actually move your work forward. ➤ 𝗣𝗿𝗲𝗰𝗶𝘀𝗲 Refine your ask. Mention your audience, desired format (bullets, list, outline), or even tone. Think briefing, not brainstorming. ➤ 𝗧𝗲𝘀𝘁 & 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗲 Great prompts don’t always happen on the first try. Test variations. See what works best. AI is only as good as the human behind the keyboard. If you’re building in AI, leading a team, or just trying to get sharper thinking from your tools, PROMPT is a cheat code. 📌 Save this. Share it with your team. And let me know your go-to prompt move. 📩 DM me to learn about my AI Training on Leadership. #JeffEyet #TheBerkeleyInnovationGroup #AI #PromptEngineering #Strategy #Entrepreneurship #Productivity #Growth