Hey there,
It is Matthew Larsen, back with my newsletter.
Over the past eight weeks, our close rate went up 22% company-wide.
And that's 22% on total calls booked, not just calls that showed up. Since we sell one thing at one price, revenue per live call went up by the same amount.
Some context. Outside of Lead Generation Academy, I'm a partner in a pay per lead agency. We sell leads to 12 verticals, heavily in financial services, insurance, lending, and medical. We have about 15 sales reps, all in person in a call center, and we book about 80 calls per day: roughly 70 from paid ads and 10 from cold outreach and other sources, mostly targeted manual prospecting.
We never had a close rate problem. This is a 22% jump on top of a sales team that was already top 1% or better.
We did it with five workflows. Some are fully automated, some are part AI and part manual.
First, why AI feels like a shiny object
There are three steps to doing anything:
Learn how to do it
Create the workflow
Automate it with AI (or any tool)
If you skip the first two, AI doesn't help. At best, you'll do the wrong things more efficiently. Learning should be your number one priority.
Here's how I use AI:
Replace low-skilled, low-stakes, repetitive work. Then roll those salaries into A players. If I can automate the tasks of five people making $4k a month, I can hire one $20k-a-month person who will actually help us grow.
Run workflows that used to be impractical. Usually a mix of AI and manual.
Stack small improvements. A 1-5% gain is usually too minor to be worth the cost. But stack 30 of them and it completely transforms your business.
What I never use it for: original or creative thinking, writing, decisions, or talking to clients, unless it's template-based.
1. Appointment setting
The drop-off between application and booked call has always been our number one problem. When I joined, fewer than 30% of applicants booked a call. Three out of ten. We got that to about 50% manually over two to three years. With AI, it's now about 65%.
Part A is for people who apply but don't book. We wait five minutes to give them a chance to book on their own. Then:
They instantly get an iMessage, WhatsApp, email, and LinkedIn connection request from their assigned setter.
Clay enriches them to find their X, Instagram, Facebook, and YouTube. AI follows and messages them there too.
The setter calls them. That part is still 100% manual. We tried AI for it, and it didn't do well for us.
Two hours later, a LinkedIn request and message from our CRO. Twenty-two hours later, from my partner. Seven days later, from me.
The piece most people miss: someone on our team creates 20 LinkedIn posts per day. AI rewrites each one to be about 5% different, staggers them, and posts them on every setter's profile so they look legitimate, not like blank accounts.
Every message and action is slightly randomized. We had trouble early on with LinkedIn's automation detection and restricted profiles, but we've largely gotten past it.
Part B is for people who book. Same idea, but everything comes from their matched closer, including a selfie video (all 15 closers filmed one), a one-page screenshot of everything they need to know, and a 13-minute Loom of me explaining the offer and the unique mechanism.
I don't think this workflow is possible without AI. Speed to lead, original content for every setter, and managing all those LinkedIn accounts is too much to do by hand.
2. Pre-call research
Getting closers to research before a call is a complete grind. But when you prepare, you look 100x more professional, and you come across better because you're confident.
AI builds a one-page Google Doc for every prospect:
Business overview
Their application answers
The five most relevant client wins
The three most relevant case studies
Five personalized quick wins, pulled from a bank of about 20
The closer still spends 10-15 minutes checking it. Allegedly.
Prospects can't book less than 24 hours out, so this has time to run. Thirty minutes before the call, they get an SMS, WhatsApp, LinkedIn message, and email: "Just finished the research for your company. Excited to talk in 30 minutes. I'm bringing X, Y, and Z."
It's been really strong for show rate, close rate, and call quality.
3. Ultra-personalized follow-up
Follow-up is arguably the most important part of sales. It's also another grind. Now it's entirely automated. Our reps don't take a single action. We run a one-call close and don't call in follow-up.
I wrote the templates. They connect to banks of FAQ answers, case studies, and activation points (a tip they try in their business, it works, and they think, "I should hire these people"). AI personalizes each message using the application answers and the call transcript:
Day 1: best case study
Day 2: most relevant case study
Day 3: best activation point
Days 5, 7, and 9: answers to the three biggest FAQs from their call
Then Mondays (case study) and Thursdays (activation point) for 12 weeks
The FAQ emails used to be static: the three biggest objections in general. Now they're the ones this prospect actually raised.
4. Script adherence leaderboard
Your script exists because it's the most likely path to closing. But reps do whatever they want, and most sales managers spend their time watching them and yelling about it.
Our process is built on Jordan Belfort's Straight Line Selling System. AI scores every call on how well the rep followed the checklist, in order.
Adherence went from about 55% to about 90%.
The leaderboard puts close rate, revenue per live call, commissions, and total calls on one side, and adherence on the other. Across all 15 reps, commission tracks adherence with no exceptions. That's the best proof I can give my team that the process works.
5. Objections leaderboard
AI analyzes every transcript and shows the most common objections across the team on a rolling 30-day basis. We've only had it a month, and there aren't many surprises after years at scale, but the market shifts.
Our CRO runs a daily sales meeting where the team shares and roleplays the best responses. The top objections then go back into retargeting ads, earlier in the VSL, higher on the landing page, earlier in email flows and the sales presentation, and into the objection checklist.
We ran workflows 4 and 5 before AI. They work with one to three closers, then become impractical unless someone reviews every call full-time. Now they're back.
The takeaways
Learn the process and build the workflow before you automate.
Use AI for what used to be impractical: multi-channel speed to lead, research on every prospect, follow-up personalized to every call.
Automate the grind your reps avoid.
Replace low-skill roles and reinvest in A players.
Stack small wins. Thirty 1-5% improvements add up to a different business.
I walk through the full system in my latest YouTube video:
Watch it here: https://www.youtube.com/watch?v=zHH234jW7jU
Hope this email was worth your time,
Matthew Larsen
PS. Work with me in my Skool group at https://skool.com/leadgenerationacademy