Skip to content

Weekly outreach experiments with one angle test

Runs outreach as a weekly experiment. It sources and grades candidates, drafts one message per person with one of three randomly assigned angles, and measures which angle gets replies. A named person approves every draft and sends it from their own account.

SalesWorks with Claude Code · Cursor+1Community example
Build your promptSee requirements

What it does

  • Snapshots the week's thesis: signals, exclusions and the one hypothesis under test
  • Pulls candidates from your sources and grades each A, B or C with evidence
  • Drafts one message per selected candidate, varying only the angle paragraph
  • Ingests replies and reports response rate per angle with confidence intervals
  • Proposes next week's hypothesis and a monthly retune of signal weights

Your approval boundaries

Approval required

These actions always wait for your yes:

  • Approving or rejecting each draft
  • Sending any message, which the person does themselves
  • Applying new signal weights
  • Starting a new run before the last one is measured

You decide what happens next.

More about this workflow

This setup is for founders and small sales teams who want more time for real conversations. It takes a repeat chore off your plate and hands back something you can check quickly.

First, it snapshots the week's thesis: signals, exclusions and the one hypothesis under test. Next, it pulls candidates from your sources and grades each A, B or C with evidence. Then it drafts one message per selected candidate, varying only the angle paragraph. After that, it ingests replies and reports response rate per angle with confidence intervals. Finally, it proposes next week's hypothesis and a monthly retune of signal weights. It works with your CRM, email and spreadsheets, each on its own login with limited access.

You get the prep work done and waiting for you. The agent stops and asks for your yes before approving or rejecting each draft, sending any message, which the person does themselves, applying new signal weights and starting a new run before the last one is measured, and says plainly if a step cannot be undone.

“It never sends anything.”
Fatjon Kalemaj fatjon-kalemaj on GitHub · Oct 30, 2026
Before you startCRM · Email accessScoped permissionsClaude Code or Cursor
Explore required tools and model recommendations

Tools and access

  • HubSpot CRM

    Handles the crm part of this job. Grant read scopes, plus notes write.

    API required: Yes. Needs a HubSpot private app or OAuth app with the CRM scopes you choose.

    Sign up: HubSpot CRM

  • Gmail

    Handles the email part of this job. Grant read-only plus compose (drafts) scopes.

    API required: Yes. Needs a Google Cloud project with an OAuth app and the Gmail API turned on. The agent signs in through OAuth with only the scopes you pick.

    Sign up: Google Workspace

  • Google Sheets

    Handles the spreadsheets part of this job. Grant access to the single sheet it needs.

    API required: Yes. Needs a Google Cloud OAuth app or a service account with the Sheets API turned on.

    Sign up: Google Workspace

Official links only. Not sponsored, no affiliate links.

Model recommendations

Best pick

Claude Sonnet 5.5

Good balance of speed and quality for briefs and drafts.

$2 per million input tokens, $10 per million output tokens.

Anthropic docs as of October 11, 2026

Cheaper backup

GPT-6 Luna

Very low cost per token for simple research summaries.

$0.10 per million input tokens, $0.50 per million output tokens, for prompts up to 272K tokens.

OpenAI docs as of October 11, 2026

Our own picks. Not sponsored, no affiliate links.

Build your prompt

Add your details, then copy the prompt into your tool.

1Your details

Fill these in and the prompt updates as you type.

roles, company size, signals

2Your prompt

4 of 4 still empty
Fill these in before you run it: [WHO YOU ARE LOOKING FOR: roles, company size, signals] [THREE MESSAGE ANGLES TO TEST] [EXCLUSIONS AND CONTACT CAPS] [WHO APPROVES AND SENDS] You are the Weekly outreach experiments with one angle test agent. Runs outreach as a weekly experiment. It sources and grades candidates, drafts one message per person with one of three randomly assigned angles, and measures which angle gets replies. A named person approves every draft and sends it from their own account. Use only the details filled in above. If any [BRACKETED] item is still blank or unclear, stop and ask for it first. Do not guess. What you do on your own: - Snapshots the week's thesis: signals, exclusions and the one hypothesis under test - Pulls candidates from your sources and grades each A, B or C with evidence - Drafts one message per selected candidate, varying only the angle paragraph - Ingests replies and reports response rate per angle with confidence intervals - Proposes next week's hypothesis and a monthly retune of signal weights What you must never do without a yes from a person: - Approving or rejecting each draft - Sending any message, which the person does themselves - Applying new signal weights - Starting a new run before the last one is measured Rules: - Before any step on the list above, stop and ask. Say in plain words what you want to do, who it affects, and whether it can be undone. - If something cannot be undone, say so first, in one short line. - Write drafts, not final actions. A person sends, pays, posts, or deletes. - Keep a short log of every action you took and every source you read. - If you are unsure, ask instead of guessing.
View run instructionsPaste into: Claude Code, Cursor, OpenAI Agents SDK

Your answers stay in this browser only.

Run your workflow

Connect your tools with only the access this setup needs, then follow these steps.

Steps

Do these in order. Test on old data before it touches live work.

  1. 1Connect HubSpot CRM with its own login, not yours. Grant read scopes, plus notes write.
  2. 2Connect Gmail with its own login, not yours. Grant read-only plus compose (drafts) scopes.
  3. 3Connect Google Sheets with its own login, not yours. Grant access to the single sheet it needs.
  4. 4Paste the prompt where your harness takes instructions. See the note under the Copy button.
  5. 5Do a dry run on old or test data first. Check every draft and every stop point by hand before you let it near live work.
  6. 6Read its log after the first real run. Only widen access once a week of runs looks clean.

Safeguards and trust score

The score describes this example setup’s access, approval boundaries, ability to undo changes, and scope.

Access18 / 30

Can draft things that send or spend, so it sits close to risky actions.

Human approval30 / 30

Waits for a yes on: approving or rejecting each draft, sending any message, which the person does themselves, applying new signal weights, starting a new run before the last one is measured.

Undo and logs14 / 20

Proposes changes to live records, so the setup needs a change log and a way back.

Scope16 / 20

Runs on its own login across 3 tools. Fewer connections means a smaller blast radius.

How to raise it

  • Access: Take away its ability to send, post or spend. Let it only write drafts, and have a person press send or pay from their own account +6
  • Undo and logs: Stop it from editing live records. Have it write the change as a proposal you apply, and keep a log of every change with a way to roll it back +3
  • Scope: Cut it down from 3 connected tools to the 2 it needs most, each with its own limited login instead of your full account +2

Do all of these to reach 89, up from 78.

Want help getting there? Get an AI Trust Audit from Pixeldesigns

Reviewed by Trustcase (example setup). How scoring works

Embed badge

Show this trust score on your site or README. It links back here.

Trustcase trust score badge
Markdown
[![Trustcase trust score for Weekly outreach experiments with one angle test](https://trustcase.dev/badge/weekly-outreach-experiments-with-one-angle-test.svg)](https://trustcase.dev/use-cases/weekly-outreach-experiments-with-one-angle-test)
HTML
<a href="https://trustcase.dev/use-cases/weekly-outreach-experiments-with-one-angle-test"><img src="https://trustcase.dev/badge/weekly-outreach-experiments-with-one-angle-test.svg" alt="Trustcase trust score for Weekly outreach experiments with one angle test" height="20"></a>

More sales workflows

Other workflows that use the same tools

  • Vendor quote compare

    Operations. Lines up quotes from several vendors side by side.

  • Spoken site update to daily report

    Operations. Turns a short spoken end of day update from a crew lead into a structured daily report. Plain rules, not the AI, decide who hears about each issue.

  • Review request texts after a sale

    Marketing. When a sale or service is marked done, it waits two days and texts the customer a link to leave a review, then one reminder. It checks opt outs first and tells the owner if a text fails.