Done for you cold email agency · LinkedIn

A done for you cold email agency ran its own LinkedIn outreach on Sellinger

Sales Automation Systems sells outbound to other companies. It used Sellinger's white label platform to run its own LinkedIn conversations: one AI persona per account, an opener within minutes of an accepted invitation, and a booked call handed to the agency's team. This is what the workspace shows, including that replies were easy and recorded conversions were few.

Client: Sales Automation Systems
Channel: LinkedIn only
Window: 1 May to 10 October 2026
Status: The account is no longer active on Sellinger
Data as of: 19 October 2026
Written by: Asen Levov
Reviewed by: Viktorio Shopov

people messaged on LinkedIn, first message from 1 May 2026

replied: 1,074 people wrote back after the first message

min

median time for the agent to answer a prospect, over 1,340 messages

leads recorded as converted in the CRM (a CRM field, not read thread by thread)

In short

The agents earned a lot of replies: 37.6% of people messaged wrote back, and the first message alone brought 24.3%.
Recorded conversions were low: 7 conversions and 9 meeting stage leads from 1,074 people who replied. Reply volume did not translate into recorded results.
The agent's job ended at a booked call. The agency's team ran the call and closed. No revenue is claimed.

6 min read

The key numbers, with their bases

This page covers the Sales Automation Systems workspace and nothing else. LinkedIn only. People first messaged from 1 May to 10 October 2026, the period the account ran on Sellinger. Data checked on 19 October 2026.

Sales Automation Systems workspace, LinkedIn only

Number	What it is
2,858	people messaged on LinkedIn, first message from 1 May 2026
1,074	replied, 37.6% (at least one message back after our first message)
2.4 min	median time for the agent to answer a prospect's message, over 1,340 messages
7	leads recorded as converted in the CRM
9	leads at the meeting stage or with a meeting status in the CRM

Recorded conversions are CRM fields. They have not been read thread by thread for this page.

An agency that sells outbound, running its own outbound on AI

Sales Automation Systems is a done for you cold email agency. An agency that sells pipeline needs its own pipeline, and it wanted one to one LinkedIn conversations for its own offer without an SDR team for every LinkedIn account.

It used Sellinger's white label platform, which means it ran Sellinger under its own brand. Each LinkedIn account had its own AI persona with its own voice, offer and booking link. The agents connected, opened the conversation within minutes of an accepted invitation, qualified, and handed booked calls to the agency's team. Prospects came mostly from people who engaged with posts by cold email and go to market voices.

The playbook: a specific opener within minutes, one question at a time

Described from the configuration used for the agency's accounts.

Day 0
Connection request
Usually with no note.
Minutes after acceptance
Opener
One specific fact from the prospect's profile and one question. No link.
A reply
Diagnose
One question per message. Reflect the prospect's own words. Answer price and fit questions briefly and honestly, then ask a qualifying question.
The bridge
Qualify or hand over
A binary qualifier, or a hand off to a member of the team who specialises in that situation.
Follow ups
A few short messages
Each quotes the prospect or adds something useful, then a polite closing note.
A booking
A 15 to 20 minute call
Through the persona's booking link. A person takes the call and holds the price.

Rules the personas worked by: two to four sentences, founder to founder. One question per message. Openers never carry a link.

Results

The funnel

Step	People	Share of people messaged
Messaged on LinkedIn	2,858
Replied	1,074	37.6%
Recorded as converted in the CRM	7
At the meeting stage, or with a meeting status	9

Which message earned the first reply (of the 1,074 people who replied)

Message	First replies	Share
The opener	695	64.7%
Follow up 1	244	22.7%
Follow up 2	75	7.0%
Follow up 3	32	3.0%
Follow up 4 or later	28	2.6%

About 35% of first replies came after a follow up. The first message alone brought replies from 24.3% of people messaged. With follow ups the share was 37.6%.

How many people each message reached, and how many wrote back

Message	People who reached it	Share who wrote between this message and the next
The opener	2,858	24.3%
Follow up 1	2,260	27.0%
Follow up 2	1,828	24.2%
Follow up 3	1,412	23.2%
Follow up 4 or later	1,043	25.9%

These shares count every person who wrote after each message, including people already in a running conversation. That is why they stay flat. They do not show that a later follow up works as well as the opener. Use the first reply table above for that.

min

median agent reply over 1,340 prospect messages; the slowest tenth took longer than about 4 hours

What did not work

Reply volume was strong. Recorded conversions were low. 1,074 people replied, and the CRM shows 7 conversions and 9 meeting stage leads. A 37.6% reply rate says the opener gets answered. It does not say the offer sells.
We cannot see deals. The CRM does not track deals. This page counts CRM fields and claims no revenue. A hand read of the threads may find outcomes the fields missed.
The follow up question is unanswered. Because later messages count people already in conversation, this recount cannot say which follow up earned the most new replies beyond the first reply table.
Old history does not belong in the result. The workspace holds chats imported from before it ran on Sellinger. They are left out of every number on this page.
The agency left. The account is no longer active on Sellinger, so this is a finished period, not a running campaign.

What made it work

1
Reply rate is not a result. Define what you call a conversion before you launch, and count it the same way every time.
2
The opener does most of the work. Close to two thirds of first replies answered it. About a third came after a follow up, and the share falls with each message.
3
Offer and audience decide conversion. Replies come easily to a specific opener. Meetings need an offer people want.
4
Have someone take the call and hold the price. The agent books. The agency closes.

How we measured this

Source. Sellinger production data for the Sales Automation Systems workspace, read only, checked on 19 October 2026. LinkedIn only. Email threads, a separate email campaign and any other workspace are excluded.
Cohort. People whose first AI sent LinkedIn message is on or after 1 May 2026, the period the account ran on Sellinger. Messages run to 10 October 2026. Chats imported from before that are excluded.
Messaged means received at least one LinkedIn message from the agent. Replied means wrote at least one message after an agent message.
Touch is the order of the agent's LinkedIn messages to a person. The connection note is not counted, and messages typed by a person count as touches. These by touch shares are not the same measure as the reply rate per touch in other studies, so do not compare them directly.
Speed. Time from a prospect's message to the next AI sent message. It uses only messages tagged as AI, which exist from May 2026.
Converted and meeting stage are CRM fields: status converted, or a meeting pipeline stage or meeting status. They are not hand read.
Prospect names are removed. The agency's staff and the personas are not named.