AI illustrationAI sales software · LinkedIn
We sell Sellinger with Sellinger
Since September 2025 our own AI agents have prospected for us on LinkedIn: finding people who engage with sales-automation content, opening a conversation, answering within minutes and booking 15-minute showcases. This is what a year of our own data says, including what did not work.
- Client
- Sellinger
- Audience
- founders and revenue leaders
- Channel
- Window
- Sep 2025 to Oct 2026
- Data as of
- 9 October 2026
people messaged on LinkedIn, from up to 43 profiles
replied: 10,978 conversations
showcase meetings booked through the agents
median reply to a prospect since May 2026 (70% within 5 min)
In short
- Most prospects came from an intent signal: they had liked or commented on a post about sales automation, AI SDRs or go-to-market. In October 2025, intent-signal lists replied at 55% and booked four times as often as untagged lists.
- Replies held at about 40% all year. Meetings per person messaged peaked at 2.4% in November 2025 and drifted below 1% from June 2026.
- Follow-ups added 45% more replies. Cutting reply time from 19 to 2 minutes did not lift meetings: the list was the bigger lever.
8 min read
We are our own first customer
Since September 2025 our own AI agents have prospected for Sellinger on LinkedIn. They find people who engage with sales-automation content, open a conversation, answer within minutes and book 15-minute showcases.
We wrote to founders, heads of growth, sales and revenue leaders, and agency owners. Individual contributors and about 60 competing vendors are excluded. The messages went out from up to 43 LinkedIn profiles of real people: team members, ambassadors and students. The agent writes in each person’s name and voice.
What we asked for was a 15-minute showcase. From July 2026, part of the traffic received a free-trial link instead, which converts outside these numbers. What follows is what a year of our own data says, including what did not work.
AI illustrationSignal
Intent, not lists
People who just engaged with a post about sales automation or AI SDRs.
AI illustrationWho
Founders and revenue leaders
Decision makers in growth and revenue roles, filtered by role and region.
AI illustrationHow
A 15-minute showcase
An observation opener, up to three follow-ups, and a one-line answer to “what is this?”.
The playbook: find people already thinking about outbound, then start a real conversation
Most prospects came from an intent signal: they had liked or commented on a post about sales automation, AI SDRs or go-to-market. The agents connected, opened with no pitch and no link, and followed up every 5 days.
Signal
Engagement with a post
Sales-tech and go-to-market content, filtered to decision makers in growth and revenue roles.
Day 0
Connection request
68,818 sent; 41% accepted.
On accept
Opener
One observation about their work. No pitch, no link, no vendor name.
+5 days
Up to 3 follow-ups
A new angle each time, then a clean exit line.
“What is this?”
One line, then 15 minutes
Name Sellinger, tie it to their words, offer a short showcase.
Booked
Showcase call
A Sellinger person runs a 15-minute demo.
Replies held at about 40% all year. Meetings per person peaked early, then fell.
Reply rates stayed between 35% and 45% for twelve straight monthly cohorts. Meetings per person messaged peaked at 2.4% in November 2025 and drifted below 1% from June 2026.
AI illustrationThe funnel
LinkedIn, Sep 2025 to Oct 2026
- Messaged27,737
- Replied10,978 (40%)
- Meetings booked351 (3.2% of replies)
- Meetings held (lower bound)121
Held is a lower bound: outcomes were recorded only for meetings booked up to May 2026, and one booking per person is counted.
Meetings booked per 100 people messaged
by month of first message
- Oct 20251.4 (48/3,407)
- Nov 20252.4 (78/3,252)
- Dec 20251.9 (30/1,577)
- Jan 20261.7 (21/1,215)
- Feb 20261.3 (36/2,783)
- Mar 20261.3 (41/3,240)
- Apr 20260.9 (34/3,733)
- May 20261.2 (24/2,027)
- Jun 20260.7 (16/2,335)
- Jul 20260.8 (20/2,435)
- Aug 20260.5 (3/633)
- Sep 20260.0 (0/728)
Reply rate
same cohorts, by month of first message
- Oct 202542% (1,414/3,407)
- Nov 202540% (1,302/3,252)
- Dec 202540% (626/1,577)
- Jan 202641% (497/1,215)
- Feb 202635% (976/2,783)
- Mar 202638% (1,223/3,240)
- Apr 202644% (1,646/3,733)
- May 202643% (876/2,027)
- Jun 202638% (897/2,335)
- Jul 202638% (915/2,435)
- Aug 202639% (247/633)
- Sep 202635% (254/728)
The agents kept replying in minutes. Over the same period the list mix and message style changed, and from July part of the traffic got a free-trial link instead of a call, which these numbers cannot see.
The signal decides who to write to. The opener decides whether they answer.
Intent-signal leads beat every other list
reply rate and meetings per person messaged
- Intent signal54.6% reply, 4.54% meetings
- Untagged list37.2% reply, 1.16% meetings
- Sales Navigator search21.7% reply, 0.71% meetings
And the source of the signal matters
| Audience whose posts they engaged with | Meetings per person messaged |
|---|---|
| AI-SDR vendors and automation creators | 2.6% to 3.7% |
| Enterprise sales thought leaders | under 0.4% |
Intent-signal and untagged lists: October 2025 cohort, same agents. Search lists: whole period. Comparisons in this section use all in-scope leads, before the exclusions in the method.
Observations out-replied questions
reply rate by opener style within the same month
- Feb 2026, observation43.5%
- Feb 2026, question33.5%
- Mar 2026, observation45.2%
- Mar 2026, question33.6%
- Jun 2026, observation51.9%
- Jun 2026, question37.3%
An observation (“Seems like your focus on scaling outreach is making a tangible difference for your clients”) beat a question, at equal or better meeting rates.
Only 5 of 30,000 openers ever mentioned the post that put the person on the list. That is the next experiment.
more meetings per person from intent-signal lists than untagged lists (Oct 2025)
reply-rate gain from an observation opener over a question
spread in meeting rate between the best and worst signal sources
of 30,000 openers mentioned the post that triggered the signal
Follow-ups added 45% more replies. Speed was never the constraint.
The first message alone got 26.5% of people to reply; with follow-ups, 38.4%. The agents’ reply time fell from about 19 minutes to about 2 over the year, yet meetings per conversation did not rise with it: answering fast is table stakes, not the lever.
Reply rate at each touch
of people who got that message without having replied yet
- Intro26.5% (7,119/26,904)
- Follow-up 19.8% (1,739/17,774)
- Follow-up 26.8% (949/14,014)
- Follow-up 35.6% (435/7,768)
- Follow-up 44.7% (69/1,485)
Share of first replies that were negative
| Intro | Follow-up 1 | Follow-up 2 | Follow-up 3+ |
|---|---|---|---|
| 0.9% | 8.5% | 14.3% | 14.7% |
median reply since May 2026; 90% within an hour
of the 48% of messages arriving at night or weekends were answered before the next morning
of conversions came from a reply to the first message
of bookings confirmed in the chat needed at least one more nudge after the link
of replied threads since May ran with no human message
median from first message to first reply
Answering in two minutes is table stakes. The list is the bigger lever.
Inside the conversations
Three real threads, condensed. Prospect names are changed; agent names are Sellinger team profiles.
AI illustration“Sign me up. Your approach is different, I will surely learn a thing or two”
Growth data leader, Madrid
Product leader with two AI products
Won after two silent follow-ups: booked on day 42
- Georgi (agent)Day 0Hey Aram. Saw your work with [his two AI products]. As someone scaling multiple AI products, how are you approaching outbound for new client acquisition?
His products, one question.
- Georgi (agent)Day 11Following up on this. Is it fair to assume outbound sales isn't a focus for [his products] right now?
Follow-up 2: an easy out.
- Aram, product leaderDay 13I would be interested to know what you offer as part of outbound sales? [...] The target are OTEs, Hotel website, trip planners, etc.
He asks to be sold to.
- Georgi (agent)+3 minGood question. It's a system where AI agents manage the full outreach process, engaging leads like the OTEs and trip planners you're targeting. The system is much easier to show than to explain over chat. Are you open to a quick 15-min showcase?
His buyer list in the answer.
- Georgi (agent)Day 40No worries if things are busy on your end. Just wanted to see if that 15-min showcase is still something you'd like to explore [...]
Fourth nudge, a clean way to say no.
- Aram, product leaderDay 42Sorry for my silence. I had an accident and was not working. Just scheduled the call
Booked; demo held, then he signed up.
Consultant who built his own system
Two objections, and won back after we were late
- Kieran, consultantDay 0I've built my own system with n8n and we've also trialled an external provider for 3 months - it didn't perform well given the volume of emails it sent. Do you have the stats / data to show?
Objection: built his own, vendor trial failed.
- Agent+6 minThat's a fair question. Building your own system with n8n puts you ahead of 99% of the market. [...] The volume-based systems you've trialled are designed to scale automation, which erodes trust. We scale trust first.
Credits what he built, reframes.
- Kieran, consultant+8 minFair enough- Happy to put in 15 minutes next Friday between 3-5pm
Booked after one nudge.
- Kieran, consultantDay 15Hey mate - if you're late to a call, particularly when it's 15 minutes then I'm not interested
Our host joined 15 minutes late.
- Agent+1 hKieran, you're absolutely right. My apologies for being late and wasting your time. I respect your decision.
One honest sentence, no excuses.
- Kieran, consultant+26 minOkay, well honesty is important too and I respect your reply. Let's book something in for next week
Won back; he rebooked himself.
Automation consultant
A discovery conversation the prospect called useful, and a rebooking 20 days after we let go
- Tsanka (agent)Day 0Niko, I was looking at your work in AI and NoCode for business automation. Seems like your focus on scaling outreach and cost savings is making a tangible difference for your clients.
Observation opener, no question.
- Niko, consultant+40 minI was waiting for your message after we connected. Yes, I help different companies optimize their processes [...]
- Tsanka (agent)+3 hThat's a really comprehensive system you've built, especially the self-improving prompts idea. [...] That's exactly what Sellinger does, but as a dedicated solution. Might be worth a quick 15-min look?
After real discovery, names Sellinger.
- Niko, consultant+2 minHonestly, thanks for these questions they really helped me see my own approach from the outside. [...] are you a recruiter, or a business owner looking for an expert to work with?
- Tsanka (agent)+2 minHonestly, my questions are tied to the Sellinger system I just mentioned. Still, would be good to show you how it works in 15 mins.
A straight answer. Booked 9 minutes later.
- Tsanka (agent)Day 1No worries at all, Niko. Sounds like you're dealing with something critical. Feel free to grab another slot [...]
He missed it for a client emergency. No guilt.
- Niko, consultantDay 36Hi! I booked a call there for today. I hope this is still relevant considering how long I've been ignoring you?
Rebooked on his own, 20 days after the exit line.
“respect the persistency, do you have a calendar link for a 15 minute chat”
“so far Sellinger has been great.”
What made it work
- 1
Write to people already reading about the problem. In October 2025, intent-signal lists replied at 55% and booked four times as often as untagged lists.
- 2
Pick the right signal source. Audiences of AI-SDR and automation content booked; audiences of enterprise sales voices barely did.
- 3
Open with an observation, not a question. Observations out-replied questions by 10 to 15 points.
- 4
When asked “what is this?”, answer in one line. Then offer 15 minutes. The booking link typically went out after three prospect messages.
- 5
Nudge after the link. More than half of bookings needed one.
- 6
Speed is table stakes. Cutting reply time from 19 to 2 minutes did not lift meetings. The list was the bigger lever.
- 7
Show up. The meeting is where the human side has to be as reliable as the agent.
How we measured this
- Source: Sellinger production data for our own workspace, read-only, snapshot 9 October 2026. LinkedIn prospect outreach only; demo accounts, onboarding emails, inbound mail and live chat are excluded.
- Excluded so the numbers are fair: every lead from one profile that ran a research-style opener we have since retired, and every thread about a different product (an integrity-screening tool) pitched from the same workspace. 2,569 people in total.
- Messaged = received at least one message after connecting. Replied = wrote back after our first message.
- Meetings = one booking per person through the agents’ booking link (a lower bound). Held = recorded as completed; outcomes stopped being recorded in June 2026.
- Response time = prospect message to the agent’s next message; human-typed replies excluded.
- Prospect names are changed. Agent names are Sellinger team profiles.
- Images are AI illustrations.
Want numbers like these for your pipeline?
Book a 15-minute walkthrough of the agent, the playbook and the numbers behind these case studies.
