Good morning.
I keep getting asked which prompts I run for sales and quarterly planning. The honest answer is that I have stopped looking for new ones.
The same handful does the work every week, and there is no reason to swap them out for something untested when there is revenue on the line.
So today's issue is the nine that have earned a permanent slot in my stack. Four are for sales work. Three are for quarterly planning, which most of you are starting on for the next quarter. Two are for the harder thing, which is keeping your own thinking sharp.
Each prompt comes with what to load before you run it and which model to use. Both matter as much as the prompt itself.
— Sam
IN TODAY’S ISSUE 🤖
Why I Still Run These By Hand
Four Sales Prompts That Tell The Truth
Three Pressure Tests For Your Planning
Two Self-Checks I Will Not Automate
Why The Prompt Is The Smallest Part
Let’s get into it.
Why I Still Run These Prompts Myself
You can run any of these on a schedule. Or you can hand them to an intelligence agent that fires every Monday and use the quarterly ones to a planning agent that runs on the first day of the quarter.
Most operators who have built their cognitive folder should be moving toward this kind of automation. Recurring intelligence is what the agent layer is for.
There is a different reason to run them by hand. The reason is you.
The work of pointing the model at your real context, reading what comes back, pushing on it, refusing the easy answer, asking the follow-up the model did not anticipate, that is thinking.
The agent version off-loads the answer to the system. The manual version upgrades the operator running the system. Those are different products of the same prompt.
The nine below are tools for both modes. Run them as agent jobs when you want recurring intelligence with no repetition tax or run them as conversations when you want to sharpen how you see your own business.
If you have to pick one, pick the conversation. Agents make your business run. Conversations make you the kind of operator who can read what your business is showing you.
1. The Pre-Call Stack
For any call where the meeting will turn on whether the prospect feels read.
Here is everything I have on this account: the last three emails between us, the most recent CRM notes, the relevant ICP file from my cognitive folder, and the offer doc that fits their stage. Read all of it. Then give me five things. What they want from this call, which may not match the agenda they sent. The question they are afraid to ask. The part of my offer that maps directly to their stated problem. The part that will not land yet, and why. The single move I should make in the first five minutes that earns the right to ask harder questions later.Context to load: cognitive folder (ICPs, offers, voice), CRM notes, recent email threads, the prospect's website and LinkedIn if you have time. Skip the company press releases. They are noise.
Best model: Claude Sonnet 4.7 with the cognitive folder loaded as a project. Claude Opus 4.7 if the deal is large enough to justify the extra reasoning depth.
What you walk into the call with: the question the prospect is afraid to ask. Build the call around that one moment and everything else falls into place.
2. The Pipeline Triage
Run this every Monday morning before your first call.
Here is my full pipeline as of today. Each row is a deal: name, stage, dollar value, last activity date, last activity content, owner, and any notes I have added. Bucket every deal into one of three categories. Real means the deal is moving and there is a clear next step on the calendar. Stalled means it has been in the same stage past my median cycle and nothing has happened in fourteen days. Ghost means it is still in CRM but the prospect has gone dark and nobody on my team has taken the hint yet. For the top three deals in the stalled bucket by dollar value, tell me the smallest, lowest-stakes move I could make this week to get a real signal back. Do not give me 'send a follow-up email.' Give me a specific message or call request and what I am trying to learn from it.Context to load: pipeline export (CSV is fine), your typical cycle length per stage, recent activity logs.
Best model: Claude Sonnet 4.7 or GPT-5.5 for pattern recognition over structured data.
What you can stop guessing about: which deals are alive and which ones you are still hoping for. The hope is the expensive part.
3. The Loss Pattern
Run this once a quarter, after the books close.
Here are the last fifteen deals I lost or that went dark in the past six months. Each entry has the prospect, the stage where the deal died, what they told me when they pulled out (if anything), and what I think the real reason was. Ignore my stated reasons. Look at the structural pattern across the whole set. What is killing my deals? Is the pattern stage-specific, ICP-specific, offer-specific, or about my process? Tell me the pattern, then give me the test I could run in the next thirty days to confirm it or rule it out.Context to load: lost deal records with notes, your current ICP file, your offer docs, your sales process documentation if you have it.
Best model: Claude Opus 4.7 with extended thinking. The whole point is to find a pattern you have been rationalizing away, and the reasoning depth matters.
What this exposes: the lie you have been telling yourself about why deals do not close. The prospect's stated reason is usually polite cover. The real reason shows up only when you read the losses as a dataset, not as individual stories.
4. The ICP Mirror
Run this whenever you are about to set targets for the quarter ahead.
Here are my last fifteen closed-won deals from the past two quarters. Read them as a dataset, not as individual wins. Find the patterns in who actually buys. The role of the buyer. The trigger event that started the conversation. The timeline from first touch to close. The size of company that converts fastest. The language they use when they describe what they bought from me. Then compare that to the ICP I have written down in my context file. Where am I wrong about who my customer is? Where am I leaving money on the table by chasing the wrong sub-segment?Context to load: closed-won records, the current ICP doc from your cognitive folder, sales call transcripts or testimonial language if you have it.
Best model: Claude Opus 4.7.
Where the gap shows up: between the ICP you wrote when you started the company and the customer who pays you now. Most operators are still selling to the version of their ICP from two years ago.
5. The Pre-Mortem
Run this on the first day of the quarter, before your priorities harden.
Imagine it is ninety days from now and the quarter went badly. Not catastrophically. Just clearly off-target. We hit sixty-five percent of the revenue goal. Three of the five priorities slipped. The team is frustrated and I am not sure where to put the blame. Walk me through the most likely sequence of events that got us there. What did we miss? What did we say yes to that we should have said no to? Where did execution break down? Where did the strategy turn out to be wrong? Do not be diplomatic. Tell me the failure I am most likely to walk into based on what you know about the business.Context to load: current quarter plan, last quarter's actual results, current team and capacity, the priorities you are weighing for the next ninety days.
Best model: Claude Opus 4.7 with extended thinking.
What it forces into view: the failure mode you are already uncomfortable with but have not addressed. The model will name it because it has no political reason to dance around it.
6. The Capacity Honest-Look
Run this after you write the priorities and before you commit to them.
Here is my current team, what each person owns, and the goals I have set for next quarter. Do a brutally honest capacity audit. Where am I expecting more output than the team can produce? Which person is the bottleneck on which initiative? What is a goal I have set that requires a hire, a contractor, or a serious tooling investment that I have not budgeted for? Do not give me 'you can do it with focus' answers. If the math does not work, tell me where it breaks and what the smallest unlock would be.Context to load: org chart with what each person owns, next quarter's goals, last quarter's actual output by person if you can pull it, current tooling and contractor budget.
Best model: Claude Opus 4.7.
What this catches early: the goal that is going to slip, six weeks before it slips. You can resource it now or downgrade it now, and either is better than discovering it in week ten.
7. The Quarter-End Forcing Function
Run this once you have the ICP refined, the pre-mortem run, and the capacity audited.
Help me set three things that must be true at the end of the quarter for me to call it a win. Not nice-to-haves. Specific outcomes that mean the business is in a different position than it is today. For each one, work backward. What has to be true at the thirty-day and sixty-day marks for the ninety-day target to be likely? Then tell me the single decision I should make this week to put the first thirty days on track. The decision should be uncomfortable enough that I have been avoiding it. If you give me three safe decisions I will know you held back.Context to load: current state of the business (revenue, team, top initiatives), historical quarter performance, the strategic options you have been weighing.
Best model: Claude Opus 4.7 with extended thinking.
What this pulls out of you: the decision you have been postponing because you are not sure you want the answer. That is usually the decision that defines the quarter.
8. The Decision Review
Run this on a decision you made roughly ninety days ago, especially one that has already shown its first results.
Here is a decision I made about ninety days ago: [describe what the decision was and what I chose]. Here was my reasoning at the time: [the doc you wrote, the Slack thread to the team, the email you sent, whichever artifact captured the call at the time]. Here is what has happened since: [the results, what played out, what surprised me, what did not]. Tell me what I got right. Where my reasoning was sound but the outcome was lucky. Where my reasoning was wrong but the outcome was good, which is the most dangerous case. Where I should have weighted the inputs differently. Then tell me the pattern in how I made this call, because I am going to make calls like it again.Context to load: the original decision artifact (doc, Slack thread, email, recording), the data that informed it, the outcome data, your honest read of what played out.
Best model: Claude Opus 4.7 with extended thinking.
What you'll uncover: the gap between reasoning quality and outcome quality. The two look identical when things go well, and they are not the same thing. Operators who treat them as the same get lucky for a while, then make a call that breaks something.
9. The Operator's Question
Run this when you have an hour and feel stuck without knowing what you are stuck on.
Here is everything I have on the current state of my business: [your CLAUDE.md or top-level context, your goal docs, your weekly notes, plus the things that have been bothering you for weeks without you doing anything about them and the conversations you have been postponing]. I am not asking you to solve a problem. I am asking you to find the question I am not asking. What is the most important question about this business that I am avoiding because I do not want to know the answer? What is the question I keep almost asking myself before changing the subject? Pose three candidates. Then tell me which one matters most and why.Context to load: your CLAUDE.md or cognitive folder if you have it, your last few weekly notes if you keep them, the things you have been deferring, and any honest accounting of what is making you uncomfortable that you have not put into words yet.
Best model: Claude Opus 4.7 with extended thinking. The model needs reasoning depth to find the question underneath the symptom.
What you weren't ready for: the question your gut already knows the answer to but your conscious mind keeps walking past. Most operator stagnation is one of those questions left unasked for a quarter too long.
What All Nine Have In Common
Every one of these prompts works because the context is already there. The pre-call stack runs because your ICP file, your offer docs, and your voice guide are sitting in a folder the model can read. Pipeline triage runs because your CRM exports structured data instead of twelve hand-typed status updates. The loss pattern works because you are logging lost deals with the real reason, not the polite version.
The prompt is the smallest part of any of this. The context is the asset.
If you have been getting generic answers from Claude or ChatGPT for these kinds of questions, the prompt is rarely the problem. The model has nothing to anchor to. Build the folder first. Then the nine above will work for you the way they work for me.
The most recent issue was the build for the cognitive folder these prompts assume. If you have not run that build yet, do it this week before you run any of the nine above. The folder is what makes the prompts compound.
If you want to go deeper than the prompts above, that is what Cortex is for.
Cortex is a monthly premium version of Bionic Business, helping entrepreneurs and business owners grow and scale with AI and Agents.
Some of the work in Cortex covers the agent architecture that automates the operational prompts on a recurring schedule.
The rest goes after what the closing two prompts hint at: how a model with the right context becomes the most consistent thinking partner an operator has ever had, what changes in your judgment when you run that loop weekly for a year, and what it takes to use AI to become a sharper operator instead of a faster one.
The agent layer scales the operation. The thinking layer scales you. Cortex goes deeper on both.
If you want more in-depth issues like this one, Cortex is perfect for you:
You only get access to the Cortex (Signals and Circuits) issues for the months you're subscribed to, not the archives.
You can cancel at any time. But there are no refunds, for any reason. All sales are final.
Full regular email issues
1x Signals Strategic Briefing
1x Circuits Tactical Guide
Access to all Skills issues
Claude Cowork as your marketing team special issue
50% OFF select workshop trainings
Get in now in time for the next batch of issues.
Until next time,
Sam Woods
The Editor
.
