Instinct sends five human-approved outreach emails
Chary reports that Instinct researched and personalised freelance outreach, then sent five emails after approval; the article provides no verified campaign metrics.
linkedin.com
Original prompt
Shared by the source author. See source notes for context.
Act as my SDR. For each lead, research their company and role, write a 60–90 word cold email
Reach people exploring AI agents
See placements →What happened
Chary reports that Instinct researched and personalised freelance outreach, then sent five emails after approval; the article provides no verified campaign metrics.
Reported by Sumanth CharyHow the agent did it
- Prepare lead addresses with company, role and relevant public context.
- Ask Instinct for short, problem-focused personalised drafts.
- Review the researched opening and social-proof claim before approval.
Where a person stepped in
The source confirms approval before five sends. Ten-draft language elsewhere is not reconciled, and no message log or measured campaign results is available.
Sumanth Chary describes testing Instinct for freelance-client outreach. He prepared a targeted lead list containing email addresses, company, role and recent-post context. His instruction asks the agent to research each prospect, write a short problem-focused email, offer an audit call and include a social-proof line.
Instinct uses public company copy, LinkedIn headlines and recent content to prepare personalised openings. The source's narrow execution claim is that five emails were sent after approval. That provides a more defensible outcome than its broad promises about handling the complete outreach lifecycle at scale.
There is an unresolved count difference: the workflow text mentions pausing the first ten drafts, while a caption confirms five approved sends. The article does not show an accessible message log that reconciles those numbers, and its claims of higher reply rates and saved hours have no figures. Readers should treat this as a small approved-send experiment. The source reports throttling, a separate outreach subdomain and opt-out handling, but does not establish delivered follow-ups or closed clients. The useful preparation is explicit prospect context and a review-before-send boundary; a plausible personalisation or social-proof claim still requires the sender to check that it is true.
Prompt
Original prompt
Source: LinkedIn · Sumanth Chary
Frequently asked questions
How many emails are actually confirmed?
A caption states that all five were sent after approval. A separate first-ten-drafts description is unresolved.
What grounds the personalisation?
The author supplies lead context and describes research using public website copy, LinkedIn headlines and recent content.
Did replies or sales measurably improve?
The article claims higher replies without figures. It provides no verified campaign metrics or closed-client outcome.
More details
Remaining steps
- Authorise the small batch; the source confirms five sends.
About this account
Instinct researches public prospect signals, drafts personalised outreach and sends an approved batch from the connected inbox.
Firsthand account by Sumanth Chary; original publication date verified against LinkedIn article Published Sep 21, 2026. Outcomes are the author’s reports, not independently reproduced. Original 21-word prompt excerpt; the complete instruction also discusses an audit call and sending. Five approved sends are reported, while ten-draft language is unresolved.
Credit: Sumanth Chary
Source published 21 Sept 2026
Source checked 14 Oct 2026
Tools and topics
Instinct · Connected email inbox · Public company sites · LinkedIn
Approved outreach · Prospect context · Five-email experiment
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.