Show Notes
The Smart Prompts for AI podcast just released an episode exploring my social voice system for solo operators. This is the same system I use with clients who would rather post nothing than post something that does not sound like them. If your best posts are the ones you delete, if you would rather stay quiet than publish something generic, or if you spend hours sanding AI drafts until they sound human, this breakdown is built for you. What We Cover in the Episode The hosts dig into why brand voice tools fail, and why the answer is not a better adjective: an adjective is not a voice, a voice is a set of measurements taken from writing you have already finished. Here are the primary systems we break down in the show: 1. The Voice Capture Feed the model three to five pieces you finished and published yourself, and it builds a Voice Card: how you open, your sentence rhythm, the words you reach for, the words you never use, how you make a point, and how you close. Then it has to pass a test in front of you, three lines that sound like you and three that are close but wrong, so you can correct it in plain words. A consultant fed it three client emails and two LinkedIn posts, and his next draft needed two line edits, after a brand voice tool draft the week before needed a full rewrite and got deleted. 2. The Voice Card, Installed A card sitting in one old chat is a measurement written on a napkin. Install it where you actually write, in ChatGPT Custom Instructions, as a Claude Style or inside a Project, or, in Gemini, pasted from a doc at the top of the session, so every draft starts from your measurements. A real estate operator used a compressed card for daily market notes and cut his routine from thirty minutes of heavy editing to six minutes of reading and tweaking one sentence. 3. One Week From One Thing Take one real event from your week, rough notes from a client call, a mistake you had to fix, a surprising metric, and pull five different post angles from it: the lesson, the mistake, the number, the broader question, and the receipt, one hyper specific detail that proves you did the work. The golden rule is invent nothing. If your notes do not support an angle, the prompt skips it instead of padding it with generic filler. 4. The Batch Board One master post, reshaped for every platform you use, in a single table, with character limits, formatting rules, and line breaks handled, while your voice card stays constant. Jeff Payne runs social media across 17 sites for one client. Reviewing and tweaking posts for those 17 sites took an entire afternoon. With the Batch Board, he approves one master post and reviews the reshaped set in about twenty minutes. 5. The Anti-Slop Pass Even with good measurements, drafts still pick up machine cologne, the lingering scent of AI in lines like “delve,” “unleash your potential,” or “a rich tapestry.” This editor pass hunts any line that could appear word for word on a competitor’s website, fixes it against your voice card rules, and runs a giveaway test: name the single line that sounds least like you, fix it or cut it. A coach ran it on a week’s posts, it flagged nine problematic lines, including two identical openings she had not noticed, and she shipped four posts with total confidence. 6. The Sunday Batch Run the whole pipeline as one thirty minute weekly habit: collect your rough notes, generate your angles, reshape them for your platforms, run the Anti-Slop Pass, and read every post out loud before it goes live, because your ear catches what your eyes miss. A solo founder who dreaded drafting sessions now does thirty minutes once a week, and she finally got her evenings back. Key Timestamps * 00:00 – Tuesday Morning, Coffee Cold: a three hour to twenty minute client win, and a brochure draft. * 01:13 – Why That Frustration Happens: James Sterling, Pacific Northwest consultant, prompt systems since 2019. * 02:04 – The Copy Paste Tax: forty minutes, five times a week, over three hours gone t
More from Smart Prompts For AI Podcast
The Smart Prompts for AI podcast just released an episode breaking down one of my most practical systems: They took my four hook-writing prompts for experts whose work nobody clicks, the same frameworks I use with consulting clients whose content deserved better than it got. In this Episode, we start with Mark: a commercial real estate operator who spent six months writing 2,000-word market analyses, then watched a gym selfie captioned “Real estate is a mindset” pull 4,000 likes against his three. One rewritten headline later, the same expertise drew 50 comments from business owners and two listing leads. If your best work is the stuff nobody opens, if you refuse to write clickbait but cannot afford to be ignored either, or if your headlines read like academic abstracts, this breakdown is built for you. What We Cover in the Episode The hosts dig into why expertise loses the feed to platitudes, and why the answer is not louder hype: clickbait is a lie, a hook is a promise, and professionals are allowed to make promises. Here are the primary systems we break down in the show: 1. The Curiosity Gap Generator Experts bury the lead by putting the conclusion in the headline. This prompt hunts your finished draft for its most counter-intuitive fact and tilts it into a question, with hype phrases banned outright. Mark’s sublease report became “Why 40% of Seattle’s ‘empty’ offices are actually off the market,” and his clicks tripled. 2. The Specific Benefit Refiner Adjectives are opinions; numbers are facts. This prompt strips the adjectives from your hook and replaces them with the hardest outcomes already inside your text, and where a benefit was never numbered, it leaves a placeholder instead of inventing one. The prompt that refuses to fabricate your results for you. 3. The Pattern Interrupter Brains are prediction machines, and a hook that breaks the prediction earns the half-second that matters. Four styles, from the quit-doing-that hook to the myth-buster, all leashed to one constraint: the hook must be supported by the text. The Chelsea Test: if the payoff is not in the piece, the hook does not ship. 4. The Story Opener Most expert content opens with throat clearing. This prompt finds the inciting incident in your draft and starts there, in the middle of the action, first person, no lesson yet. Advice raises the reader’s guard; a story lowers it. This episode opened inside one. Key Timestamps * 00:00 – The Lunch Table: 4,000 likes for a platitude, three for six months of expertise. * 01:50 – The Asset, Named: a paywalled post, and why the copy-paste prompts stay behind the wall. * 02:26 – Who Is James: University of Washington computer science, 2012; Pacific Northwest consultant; prompt systems since 2019. * 03:29 – The Thesis: clickbait is a lie, a hook is a promise, and Mark was refusing to promise anything. * 04:03 – Digital Numbness: the callus readers grow, and the two-second filter your headline has to survive. * 05:31 – Information Gap Theory: Loewenstein’s itch, and why the gap has to be bridgeable. * 06:49 – Prompt 1, The Curiosity Gap Generator: hunt the counter-intuitive fact; hype phrases banned. * 07:53 – Mark’s Result: the empty offices headline. Clicks tripled, 50 comments, two listing leads. * 08:30 – The Skeptic’s Objection: isn’t a hook just dressing? For written work, the hook is the distribution. * 10:22 – Prompt 2, The Specific Benefit Refiner: adjectives are opinions, numbers are facts. * 11:19 – The Placeholder Rule: when the draft never stated the number, the prompt leaves a blank instead of a hallucination. * 11:45 – The Bakery Result: from a lullaby about flour ordering to an 18% waste cut in month one. * 13:25 – Prompt 3, The Pattern Interrupter: prediction machines, and the half-second stumble. * 15:01 – The Chelsea Test: no strawmen, and promise dirty water only if the piece is about dirty water. * 15:41 – Prompt 4, The Story Opener: in media res, first person, no lesson at the top. * 16:46 – The Met
If writing the instruction blocks is the part you dread, Smart Prompt Studio turns a plain-words description of the job into a ready-to-paste agent prompt, plus three variations. Five free generations, no card: smartpromptstudio.com/lifetime
The Smart Prompts for AI podcast just released an episode breaking down one of my most requested systems: They took my 7-step workflow for turning a dead newsletter into a growth engine, the same positioning, audience, and systems work I used to do with consulting clients. In this Episode, we start with the brutal math: most newsletters die at forty-two subscribers, and twelve of those are your relatives. Four hours on a post, a thirty percent open rate, twelve readers. Twenty minutes of your life per reader, and zero strangers reached. 👇Read the Full Article & Copy the 7 Prompts👇 https://smartpromptsforai.substack.com/p/why-nobody-is-reading-your-substack If you are publishing great work to an audience of dozens, if your About page reads like a horoscope, or if week three keeps defeating you, this breakdown is built for you. What We Cover in the Episode The hosts dig into why “write great content and the audience finds you” is the fairy tale of the modern internet, and why you have an infrastructure problem, not a writing problem. Here are the primary systems we break down in the show:
👉 Read the Full Companion Article & Copy the 4 Prompts Herehttps://smartpromptsforai.substack.com/p/how-i-use-ai-to-learn-anything-in The Sunday Tab Graveyard Most solo operators respond to an unfamiliar industry by trying to out-read the internet. You build what I call a tab graveyard: dozens of open browser tabs, unread PDFs, and academic overviews that give you zero context on how a business actually makes money. If you bill $150 an hour and spend six hours hunting through forums on your weekend, that is a $900 tax on your own schedule. Worse, you walk into Monday morning feeling more disoriented than when you started. Marcus, a boutique strategy consultant out of Atlanta, told me recently: “I used to spend eight hours studying trade publications before an intro call, only to freeze the second the prospect asked how their vendor pricing compared to market standard.” The problem is never a lack of reading speed. The problem is treating an AI model like a search engine instead of using it as a private research analyst. You do not need to read harder; you need to build the map before you look through the microscope. What We Cover in Episode 50 * The Tab Graveyard Trap: Why speed-reading whitepapers burns billable time and guarantees you retain nothing useful when the Zoom call goes live. * Map Before Microscope: How to command the AI to lay out the commercial landscape, foundational pillars, and plain-English analogies before touching a single technical detail. * The Context Injection Trick: How stating your precise consulting objective steers the model away from generic corporate summaries and directly into high-leverage commercial answers. (If you want to skip prompt engineering fatigue entirely, grab lifetime access to Smart Prompt Studio (https://www.smartpromptstudio.com/lifetime) to build these prompts in seconds). * Following the Capital: The exact prompts to map market incumbents, venture funding velocity, and the scrappy startups threatening legacy players. * Mining the Fault Lines: How to uncover the internal debates, industry skepticism, and overhyped trends that instantly separate an experienced operator from an outside tourist. * Pre-Mortem Sparring: Setting up the model as a skeptical founder to grill your questions and provide live coaching notes before you ever risk a real conversation. * The Live-Fire Safety Net: A simple two-sentence protocol to handle dense technical jargon during a live client call without missing a beat.
Spend 3 minutes on the Reader Calibration Survey to shape what we build next. Be one of the first 100 to lock in your shot at the $50 gift card.