Your BDRs should spend their day talking to buyers, not clearing an inbox and patching the CRM. This is the exact Cowork setup that drafts inbound replies, first touches, CRM updates and account research on a schedule, plus how JayOh rolls it out with guardrails your RevOps team can live with.
Based on Anthropic's How Anthropic's business development team uses Claude to run inbound and outbound at scale. Rebuilt and extended by JayOh, an Anthropic partner.
A skill grounded in one sales knowledge base and each rep's voice. It drafts, flags what the KB can't answer, and never sends.
A pipeline scanner compares stage rules to email, calendar and call activity, then proposes updates a rep approves or dismisses.
A scheduled sweep leaves a brief, a fit score and a next play on every account before the rep logs in.
Discovery calls scored against your playbook, and AE questions answered with a dashboard instead of a data-team ticket.
11 steps. Use Next, the dots or your arrow keys. Orange boxes point at what matters on each screen; a few buttons actually work.
Tip: tap the expand button (top right of the guide) and turn your phone sideways for the full screen.
This walkthrough is an illustrative recreation of the Claude desktop app (Cowork) built by JayOh, not the live product. Every company, person, email and number is fictional.
The job is coverage. A BDR team is judged on how fast inbound gets a real answer and how many accounts in the book get worked with something relevant to say. This play moves the repeatable 70% of that work (inbox replies, first touches, CRM upkeep, account research, call review) into Cowork skills that run on a schedule and leave drafts and proposals for a rep to approve. The outcome: sub-hour inbound response, a researched brief on every account every morning, a CRM that tracks reality, and reps who spend their hours in conversations.
The BDR manager owns the skills and the knowledge base. RevOps owns connectors, CRM permissions and the stage rules. Each rep owns their own voice profile and every send.
Claude Cowork (skills, scheduled tasks, artifacts), Gmail, Google Calendar, Google Drive, Salesforce or HubSpot, Gong, and your warehouse (BigQuery, Snowflake) for usage data.
Claude Team or Enterprise seats with Cowork, connectors approved by an admin, a written stage-definition doc, an ICP doc, and 20–50 of your best past replies and outbound emails.
| Metric | Type | Weak | Acceptable | Strong |
|---|---|---|---|---|
| Inbound first-response time (business hours) | Leading | > 4 hrs | 1–4 hrs | < 1 hr |
| Drafts sent with light or no edits | Leading | < 50% | 50–75% | > 75% |
| Drafts flagged for a KB gap | Leading | > 25% | 10–25% | < 10% |
| CRM proposals approved | Leading | < 60% | 60–85% | > 85% |
| Book accounts with a brief under 7 days old | Leading | < 50% | 50–90% | > 90% |
| Rep hours per day on inbox + CRM admin | Lagging | > 3 hrs | 1.5–3 hrs | < 1.5 hrs |
| Meetings booked per BDR per month vs. pre-rollout baseline | Lagging | flat | +10–25% | > +25% |
JayOh working ranges, not industry benchmarks. The hourly schedule makes sub-hour response achievable; approval and edit rates are where we see quality settle after two to three weeks of feedback written back into the skills. Measure against your own 30-day baseline.
Not in this setup. The inbox and first-touch skills create Gmail drafts only. A rep reads, edits and sends every one. We keep it that way even after quality is high, because the send is where accountability lives.
A single sales knowledge base of common questions and approved answers, a short reply-rules doc (tone, routing, what never to answer), and a voice profile per rep built from their own sent emails. Without the knowledge base, drafts drift into guesses.
It can, but we recommend proposals first. The pipeline scanner shows each change with the email, meeting or call evidence behind it, and the rep approves or dismisses. Dismiss reasons are written back into the skill so the same mistake doesn't repeat.
Each skill runs on a schedule you set in Cowork's Scheduled view: hourly for inbox drafts, every few hours for new leads, nightly for book research, daily for the pipeline scanner. Results are waiting as task outputs when the rep opens the app.
For a team of 5–15 reps with clean CRM access, JayOh typically runs a two-week pilot with two reps, then a two-week rollout to the team. The knowledge base and stage rules are usually the longest pole, not the skills.
Leading: first-response time, the share of drafts sent with light edits, KB-gap flags and CRM proposal approval rate. Lagging: rep hours on admin and meetings booked per BDR against a 30-day baseline.
JayOh builds the knowledge base, skills, schedules and CRM guardrails, pilots with two reps, and hands your team a setup that runs without us. Talk to JayOh.
Talk to JayOh
We engineer growth systems. JayOh builds the RevOps architecture, GTM execution and data infrastructure that turn messy revenue operations into predictable pipeline for B2B SaaS and PE-backed teams.
