Today’s plays › All plays › Prospecting & research
Have AI read your whole LinkedIn message history to find what gets replies, and the yeses you never answered
Picked and checked by Olga Pearson, SalesUpLevel
Get tomorrow’s play in your inbox. Free ↓
Reported by Peter Witte on LinkedIn: Claude reviewed over 32,000 of his LinkedIn messages. Personal story openers had a 46% reply rate, versus 0 to 9% for generic brochure-style pitches. It found several dozen neutral or positive replies he never answered, and his reply rate after 6+ steps was about double his reply rate on first contact.
Your LinkedIn inbox is a dataset you already own. Most of us never look at it as one.
Peter Witte says LinkedIn has been his number one source of closed deals for most of this year. He asked Claude to review over 32,000 of his LinkedIn messages and answer four questions: which demographics he did best and worst with, which messages converted higher, which gaps in his outreach needed fixing right away, and what those gaps cost him. Three findings came back: personal story openers that referenced his own experience beat generic pitches by a wide margin, several dozen prospects had replied neutrally or positively and never heard back, and persistence paid, with replies after six or more steps at about double the first-contact rate. He calls his analysis basic, but actionable and an easy fix.
By Peter Witte, EVP of Sales, 360 Universal Solutions (merchant services), per his LinkedIn profile · LinkedIn post ↗ · Aug 4, 2026
Run it before your 10am
- Download your LinkedIn message history (LinkedIn lets you request a copy of your data, including messages).
- Remove anything you should not put in an AI tool, and use the tool your company approves.
- Ask the four questions Peter asked (prompt below), and make it show the counts behind each claim.
- Before anything else, answer every neutral or positive reply that never got a response.
- Rewrite your opener as a short personal story that connects to the buyer, and test it against your current one.
- Keep following up past the first message, and check your own reply rate by step every month.
Attached is my LinkedIn message history from [dates]. I sell [what] to [who]. 1) Which types of people (title, industry, company size) replied most, and which did I struggle with? 2) Group my first messages by style. Give the reply rate for each style, with two examples of each. 3) List every conversation where the prospect replied neutrally or positively and I never answered: name, date, and their last message. 4) What gaps in my outreach should I fix right away, and what did each gap likely cost me? Show the counts behind every claim. If a group is too small to trust, say so.
If you lead a team
- Have each seller run this once a quarter and share one finding in the team meeting.
- Compare opener styles across the team, not just per seller.
Watch out for
- One seller's own numbers, not a controlled test. The post does not say how many messages were in each group.
- Your inbox holds other people's personal data; check your company's rules before uploading it.
- Check the unanswered list by hand before you reply; AI can misread a polite no as a maybe.
- What I'd have my team try first: step 4 alone. Answer every yes you left hanging this week.
Source: Peter Witte's LinkedIn post (Aug 4, 2026). Title from his LinkedIn profile; the post itself does not name his company. Results are as reported by him; we did not measure them.
Get tomorrow’s play before your first call
Free. One sourced play a day.