What You're Actually Building
This isn't another dashboard template. It's a system with four parts:
Today's Briefs. Every morning, Claude pulls your sales calls from your calendar, matches each lead to the application they filled out, then researches them online: LinkedIn, Instagram, X, their website, review sites. You get a 90-second brief per lead with a deal heat score, one opening play to build rapport fast, and one thing to watch out for.
Call Intelligence. Claude reads your call transcripts and shows you your close rate, your talk ratio, which objections kill you most, and the patterns behind your wins and losses. Real numbers from your real calls. Plus a breakdown of every call: summary, what you did well, what to fix, and a notes box.
Your Script.Your exact sales script lives in the dashboard, and every recorded call gets graded against it section by section. You'll see exactly which parts you nail and which parts you skip when the call runs long.
The 7am routine. All of it runs automatically as a scheduled task inside Claude. Not a cron job. Not some script duct-taped to your laptop. It just runs.
What You Need Before You Start
Three things:
1. A Claude account with the desktop app or claude.ai. The paid plan, because connectors and scheduled tasks live there.
2. Your tools.A call recorder (Fathom, Fireflies, Otter, Granola), a calendar (Google Calendar, Outlook), and wherever leads apply before booking (GoHighLevel, Typeform, HubSpot, Google Forms). Don't have all three? Start anyway. The prompt handles gaps.
3. The prompt.It's at the bottom of this page. Free.
Connect Your Software to Claude
This is where most people expect it to get technical. It doesn't. No API keys. No .env files. No code.
Open Claude, go to Settings, then Connectors. You'll see a directory of tools Claude can plug into directly: Fathom, Google Calendar, Google Sheets, HubSpot, and a growing list of others. Click connect, log in to the tool, done. Claude can now read your recordings, your calendar, your sheets.
Connect the ones you actually use before pasting the prompt. It saves a back-and-forth later.
If your tool isn't in the directory:some tools (like GoHighLevel) don't have an official connector yet. Two options. One, add it as a custom connector if the tool publishes an MCP server URL. Two, let Claude give you the workaround when you run the prompt, usually a Google Sheet or an export you drop into the chat. The prompt is built to tell you straight instead of faking a connection.
Paste the Prompt
Copy the full prompt at the bottom of this page and paste it into a new Claude conversation. That's it. Claude takes over and interviews you one question at a time: what you sell, which recorder you use, which calendar, where leads apply, whether you have a script, how you track closes, and what time you want your morning briefing.
Answer honestly, especially the closes question. If you track wins and losses in a spreadsheet, Claude connects to it. If you track nothing, Claude sets up a simple Google Sheet tracker and reminds you to log outcomes after each call. Without outcome data, there's no close rate, no win patterns, no coaching. It's the difference between a dashboard and a mirror.
Make Claude Prove Every Connection
The prompt forces this, but know why it matters: after each connector is live, Claude immediately pulls one real record. Your latest recording. Today's calendar. Your most recent form submission. You see the actual data before anything gets built on top of it.
Anything that can't connect yet gets built with sample data clearly labeled SAMPLE, so you can see the vision and swap it out the moment the connection is live. Nothing fake ships silently. That rule is in the prompt on purpose.
Turn On the 7am Routine
The last thing Claude does is create the scheduled task: every weekday at your chosen time, it pulls your calls, matches applications, researches every lead, and has your briefs waiting before your coffee.
The prompt makes Claude run the routine once manually, end to end, and show you a real brief before you call it done. If the test brief looks good, you're live. Tomorrow morning it happens without you.
What Lands in Your Dashboard Every Morning
For each booked call: who they are, what their application says (revenue, bottleneck, what they tried before, budget), what Claude found online with a source for every single claim, a deal heat score, one opening play, and one red flag.
One rule baked into the prompt: Claude never invents facts about a lead.If it finds nothing, it says "nothing found." A wrong "fact" on a sales call is worse than no fact.
Copy This Into Claude
Select everything in the box below, copy it, and paste it into a new Claude conversation. Claude handles the rest.
I want you to build me a personal AI Sales Command Center. This is my one stop shop for everything sales. Before you build anything, interview me one question at a time (wait for my answer before asking the next one): 1. What do I sell, who do I sell it to, and what does a typical deal look like (price, call length, one-call close or multi-call)? 2. What call recording software do I use? (Fathom, Granola, Fireflies, Otter, Zoom, Gong, or something else. If I use none, recommend one that has a Claude connector.) 3. What calendar do my sales calls live on? (Google Calendar, Outlook, Calendly, GoHighLevel calendar, Cal.com, etc.) And how can you tell which events are sales calls vs everything else? 4. Where do my leads fill out an application before booking? (GoHighLevel form, Typeform, Jotform, Google Forms, HubSpot, or nothing yet.) What questions does the form ask? 5. Do I have a sales script or talk track? If yes, I will paste it in. If no, help me write a simple one first: opener, discovery questions, pitch, price, close. 6. How do I track whether a deal closed? Walk me through how I want wins and losses labeled: do I keep a spreadsheet (Excel or Google Sheets), a CRM, or nothing yet? If I track deals somewhere, connect to it and read outcomes from there. If I track nothing, set up a simple closes tracker for me (a Google Sheet works great) and remind me to log the outcome after each call. 7. What time do I want my morning briefing? (Default: 7:00 AM my time, every weekday.) HOW TO CONNECT MY TOOLS (do it this way, not the complicated way): - Use Claude connectors for everything. Tools like Fathom, Granola, Google Calendar, Google Sheets, HubSpot and many others connect natively through Settings then Connectors in the Claude app. No API keys, no .env files, no code for connections. - For each tool I named: check if its connector is already active in this conversation. If it is not, point me to exactly where to enable it (Settings then Connectors, or the connector directory) and wait for me to connect it. - Once connected, test it immediately by pulling one real record (my latest recording, today's calendar events, my most recent form submission). Show me what you pulled so we both know it works. - If one of my tools has no Claude connector at all, tell me straight, and give me the simplest workaround (usually a Google Sheet or an export I drop into the chat). Do not build any feature on top of a connection that failed its test. THE MORNING AUTOMATION (this is the heart of it): - Set this up as a native Claude scheduled task (a routine) that runs automatically every weekday at my chosen time. Not a cron job, not a script on my computer, nothing weird. It should live inside Claude and just run. - Every morning the routine does this: 1. Pull today's sales calls from my calendar. 2. For each booked lead, find their application answers (match by email address). 3. Research each lead online using their full name, email, and company: LinkedIn, Instagram, X, their website, review sites, news, anything public. 4. Deliver my briefs and update my dashboard so it is waiting for me when I wake up. - After creating the routine, run it once manually end to end and show me a real brief before we call it done. WHAT EACH BRIEF LOOKS LIKE (short enough to read on my phone in 90 seconds): - Who they are and what their application says: revenue, bottleneck, what they tried before, budget. - What you found online, with a source for every claim. Never invent facts about a lead. If you find nothing, say "nothing found" instead of guessing. - A deal heat score, one opening play to build rapport fast, and one thing to watch out for. THE DASHBOARD (four sections, dark theme, clean and modern): SECTION 1: TODAY'S BRIEFS This morning's pre-call research, one card per lead, exactly as described above. SECTION 2: CALL INTELLIGENCE Analyze my past call transcripts and show: - Close rate, average talk ratio (me vs prospect), average call length for wins vs losses, and script adherence percent. - An objection leaderboard: which objections come up most, and my win rate when each one appears. - Winning patterns vs losing patterns pulled from my real transcripts, with the numbers to back each one. - Three specific coaching fixes each week, with timestamps from my own calls so I can replay the exact moments. - Use my closes tracker from question 6 to know which calls were won or lost. - Below the stats, a per-call breakdown for every recorded call. Each call gets an expandable card with: · A short summary of what happened on the call. · What I did well, with timestamps. · Things to improve, with timestamps. · The outcome (won, lost, or follow-up) pulled from my closes tracker. · A notes box where I can type my own notes about that call and save them, and they stay saved when I come back. SECTION 3: MY SCRIPT My exact sales script lives here. Every recorded call gets graded against it, section by section (opener, discovery, pitch, price, close), so I can see which parts I nail and which parts I skip. Connect my weak sections to my loss patterns. SECTION 4: INTEGRATIONS What is connected, when it last synced, and my automations: the daily morning research routine, an after-call breakdown when a new recording lands, and a Friday coaching report. BUILD ORDER (so I see progress and nothing fake ships): 1. Interview me, one question at a time. 2. Connect and test each tool through its connector, one at a time. 3. Build the dashboard using my real data from those tests. If a section has no working connection yet, fill it with sample data clearly labeled SAMPLE so I can see the vision, and swap it out the moment the connection is live. 4. Create the morning routine and test-run it once end to end. 5. Show me the finished dashboard. Start with question 1.