Three plays from Anthropic's own head of sales running a 4,000-account book. Daily call prep on a schedule, the Friday forecast on autopilot, and an overnight territory scoring for every account in your book. With the paste-ready prompts to replicate each.
In May 2026, Anthropic published a blog post from Travis Bryant, their Head of US Mid-Market GTM, walking through exactly how he uses Claude Cowork on his own sales team to run a 4,000-account book. The post is genuinely good. The plays inside it are mostly free to copy if you have access to Claude Cowork or any AI assistant with similar scheduling, connector, and skills capability.
This guide breaks down the three highest-leverage plays in his post, adds the paste-ready prompts so you can replicate each, and finishes with the scoring rubric for the overnight account-ranking move he ran across all 4,000 accounts in one night.
The setup. Travis has a scheduled task in Claude Cowork that runs every morning before he sits down at his laptop. The task pulls customer spend data and pipeline status from Salesforce, looks at his calendar for the day, and generates a one-page brief for every meeting he has scheduled. By the time he opens his laptop the briefs are already there.
Why it works. Most sales reps spend 15 to 30 minutes per meeting prepping. On a calendar with 6 meetings a day that's an hour and a half a day. A scheduled Cowork task moves all of that work to overnight (or pre-coffee), so the actual workday starts with the prep already done.
What you need. Claude Cowork (or any AI with scheduled tasks + a connector to your CRM and calendar) plus the prep prompt below saved as a Skill. Run it on a daily schedule.
Save this as a Claude Skill called Daily Call Prep. Wire it to your CRM (Salesforce, HubSpot, Pipedrive) and your calendar. Schedule it to run every weekday at 6am. By the time you open your laptop, the briefs are in your inbox or a shared doc.
Copy this prompt
You are my Daily Call Prep skill. Every weekday morning, you generate a one-page brief for every customer meeting on my calendar today. Run this in order: 1. CHECK MY CALENDAR. Look at my calendar for today. Filter out internal meetings, time blocks, and personal events. Keep only external customer meetings. 2. FOR EACH MEETING, BUILD A BRIEF in this exact format: MEETING: [Customer name + meeting time] ATTENDEES: [their name + title + role; my team's attendees] THE STAGE: [pipeline stage: new logo / expansion / renewal / churn risk] LAST TOUCH: [date and one-line summary of our last interaction, pulled from CRM activity or my recent emails] THEIR SPEND TO DATE: [from CRM, year-to-date and lifetime] THE NUMBERS THAT MATTER: - [their open opportunities, dollar amounts, close dates] - [contract renewal date if applicable] - [their usage / adoption metrics if available] THE STORY SO FAR: [3 sentences. What they bought, why they bought it, what's happened since.] THEIR LIKELY GOAL FOR THIS MEETING: [your best guess based on their stage, last touch, and recent activity] MY ONE OBJECTIVE: [the single outcome I should walk out with. A yes, an expansion conversation, a renewal commitment, etc.] TWO QUESTIONS I SHOULD ASK: [tailored to their specific situation, not generic] ONE RISK I SHOULD WATCH FOR: [the realistic objection or red flag based on stage + recent data] 3. DELIVER. Post all briefs to a shared doc with today's date, OR email them to me, OR drop them in my Slack DMs. Whichever channel I've configured. Be specific. No generic language. Every brief should pass this test: does this say something an AI couldn't have said without my real CRM data? If not, rebuild it.
His quote that matters most
Travis described the shift as "Claude builds the what, I do the why." Data assembly is the part Claude handles. The strategic conversation is the part he still does. That framing applies to every play in this guide.
The setup. Every Friday, a Cowork scheduled task pulls his opportunity records and commit numbers from Salesforce, pulls notes from his internal docs, and builds the one-page forecast report leadership wants to see. The report gets posted to a shared link before their weekly forecast call. Travis told the team it saves him about 3 hours a week.
Why it works. Forecast prep is one of the highest-frequency, lowest-strategic-value tasks a sales leader does. The data exists. The format is repeatable. Leadership reads the same shape every week. It's the perfect AI workflow.
The prompt for it is essentially the call-prep prompt above, retargeted at the pipeline as a whole instead of a single customer. Save it as a Skill called Weekly Forecast. Schedule it to run every Friday at 8am. Wire it to your CRM and your internal docs (Notion, Google Drive, Confluence).
The output should land in a shared doc the same way every week so leadership knows where to find it. Same headers. Same shape. Different numbers.
This is the play in the playbook that's the genuine unlock. The kind of project that used to take a cross-functional team hours and hours, Travis ran in a single night. He gave Claude a scoring rubric, pointed it at his entire 4,000-account list, and woke up to every account scored with a written reason, deep research, and Salesforce data attached. Then he had Claude build an interactive dashboard the reps click into to see their top-ranked accounts.
You don't have 4,000 accounts. Most teams have 50 to 500. The play still applies. Run it once a quarter against your full book. Each rep walks into the quarter with a ranked list of where to spend their time.
The rubric (paste-ready, customize for your business):
1. Recent buying signal (0-10) — did they engage with content, attend a webinar, hire someone in a target role, expand into a market we serve?
2. Fit score (0-10) — do they look like our top 20 customers on size, industry, stack, and use case?
3. Stage of relationship (0-10) — new logo prospect / past customer / current expansion candidate / renewal at risk.
4. Account-level momentum (0-10) — funding round, product launch, leadership change, news mention, anything signaling motion.
5. Reachability (0-10) — do we have a champion, a warm intro, or recent contact?
Total score out of 50. Then Claude writes a 3-sentence "why this score" for each account. The reps see the score AND the rationale, so they trust the ranking.
Run it overnight on a Sunday. The reps wake up Monday to a fresh ranking. Cycle weekly or monthly depending on how fast your buying signals change.
The bigger move
Travis's three plays compound. Daily call prep frees up the morning. Weekly forecasts free up Friday. Quarterly overnight scoring frees up the strategic re-prioritization that almost no team does well. Stacked together, you've moved the entire low-leverage data-assembly side of sales to AI and freed your reps to do the only thing AI can't do yet — actually talk to customers.
The Only AI Masterclass You Need
If this guide helped, but you’re looking to go deeper, I got you!! My 30-Day Challenge takes you from saving AI tips you never use to actually building with AI, step-by-step.
I show you exactly how I automated two e-commerce brands, my social media, and most of my personal life, then hand you the agents, workflows & systems to do the same. I’m teaching you every single thing I know with one lesson and one build a day.
Join the AI Masterclass →© 2026 Mariah Brunner. All rights reserved.