How to generate bid request emails with AI

Here’s a tactical blueprint for using AI to write killer emails: Interview 5 customers. Ask what they were doing before, what they hated, and what pushed them to act. Have a notetaker in the call. Keep peeling the onion - act like a therapist and get them to talk as much as you can for the transcriber. Take the transcripts and prompt your favorite LLM as follows: “You’re a senior outbound strategist helping a founder translate raw customer interviews into a friction-first cold email. You have 5 transcripts of real buyer conversations. From these, extract the emotional spikes, recurring complaints, and phrases that signal urgency or frustration. Write a sequence of 3 cold emails that feel like they came from someone who lives in their world. It should: 1) Lead with a pain the buyer actually said, not what we think they feel 2) Be <60 words 3) Avoid adjectives, features, and startup fluff 4) End with a low-commitment ask (e.g. “Does any of that sound like your version of reality?”) Bonus: make it sound so specific it couldn’t have been written by anyone who hasn’t lived their pain.” Test them. Track responses. Keep what hits. Iterate the rest. Until your cold email mirrors a real complaint your customer muttered under their breath last week… you’re still guessing.

Most people get AI prompting completely wrong. Here's the formula I use when I'm working with AI 👇🏼 Bad prompts = generic, unusable outputs Smart prompts = customized, high converting outputs. Here's the SMART frame that transformed my AI output quality: 1. S - Set Context Most people jump straight into asking. Smart operators give AI the full picture first. Instead of: "Write me a cold email" Try: "You're writing for B2B SaaS companies targeting VP Marketing at 50-200 employee companies, focused in the healthcare sector. You're focused on building successful angles focused on pipeline attribution challenges. Tone should be direct but consultative. Avoid buzzwords at all costs, and use data from our GTM analysis as a form of social proof." 2. M - Meaningful Dialogue Don't just command AI - invite it to ask clarifying questions. Add: "Before you start writing, what additional context do you need about the audience, their pain points, or our solution to create the most relevant message?" This opens a back-and-forth that dramatically improves relevance. 3. A - Add Examples Show AI what great looks like instead of hoping it guesses correctly. Include :"Here's an example of our best-performing email: [paste example]. Notice the hook, specific pain point being addressed, and either-or CTA. Create something with a similar structure but different content." 4. R - Request Variations Never settle for the first output. AI's best ideas often come in iterations 3-5. Add: "Please provide 2-3 different versions of this email, each testing a different pain point or value proposition." T - Tweak With Specifics Give detailed feedback about what worked and what didn't. Instead of: "Make it better" Try: "The opening hook was strong, but the value proposition feels too generic. Make the middle section more specific to pipeline attribution challenges, and strengthen the social proof element." The different is dramatic. Generic prompts produce generic content that sounds like every other AI-generated email. The SMART framework produces customized content that actually converts. Most people blame AI for poor output quality when the real problem is prompt construction. AI is only as smart as the context and direction you provide it.

Most email outreach fails for one reason: It sounds exactly like what your competitors send to your prospects. Most "personalized" emails still sound generic and flat. And your prospects? They can smell the copy-paste from a mile away. So here’s how we flipped the script using real personalization with AI — and it actually works. Picture this: You want to send an email that feels like you actually know the person reading it. But you also want to scale it across dozens (or hundreds) of leads. Here’s the step-by-step playbook we use at Level: 1. Start with your CRM- Pull real data from your client or prospect profiles. 2. Use Browser AI to research- Scrape their website to find language, priorities, and cultural cues. 3. Grab their LinkedIn insights- Who are they? What are they posting? What matters to them? 4. Build a real persona- Use the "Jobs To Be Done" framework to define their motivations and pain points. 5. Map your services- Clearly list what you offer and identify case studies that match. 6. Ask ChatGPT to matchmake- Feed GPT the persona + services + case studies and ask what fits best. 7. Write the email with GPT-4.5- Give it a clear tone direction, include the inputs above, and let it draft a personalized message. The result? An email that actually feels like it was written just for them. No spray-and-pray templates. Just smart, human-centered messaging, at scale. This approach changed the game for us. It's faster, more relevant, and makes every lead feel seen. #AI #Personalization #B2BMarketing #MarketingInnovation #ChatGPT #Entrepreneurship #Leadership #AIMarketing