AI won't save a broken sales process. It'll just help you run the broken version faster.
If reps don't know what a good call looks like, automating the bad one scales the problem. Fix the process first. Then let the tools multiply it.
Small change, real lift: give your first offer and second call different names.
"Demo" tells a prospect to brace for a pitch. "A working session to map where you're leaking pipeline" is the same slot — but now it promises them something. Name the outcome, not the activity.
The reps getting real lift from AI aren't writing "better emails." They're personalizing deeper than they could by hand.
Feed Claude the account's recent news + the prospect's role, ask for one specific opening line. Generic 8% → relevant 20%+ isn't magic. It's just relevant.
Battlecards go stale the day you write them.
Faster loop: paste a competitor's latest positioning into Claude, ask how they'd attack you and how you'd respond. Living battlecard in 15 minutes. Beats the PDF nobody's opened since Q1.
The end of a first call is where most next steps quietly die.
"I'll send some info and follow up" is not a next step. A next step has a date, a calendar hold, and a reason to meet again. If you leave without one, you probably don't have a deal — you have hope.
6/ Start with 20 deals and one question. If the answer surprises you, that's the signal your team's been flying blind on something. Most are. https://t.co/Xv4QlS6qVO
2/ Start small: pull transcripts from your last 10 won and 10 lost deals. Ask "what's actually different about the ones we won?" You're not coaching one call — you're looking for the pattern across twenty.
5/ The point isn't replacing a good manager. It's that most reps aren't drowning in coaching — they're starved for it. This gives you feedback on calls that would've gone unreviewed forever.
"Follow up" logged in a CRM means nothing. It's a note to self, not a plan.
A real follow-up has a channel, a reason, and a next step the prospect actually agreed to. Everything else is just a reminder to feel busy later.
Proposals and ROI messaging eat hours and reps hate writing them.
Give Claude the deal notes and the prospect's stated pain, ask it to draft the ROI framing in their words. You edit instead of stare at a blank page. Draft in minutes, not a morning.
Underrated follow-up channel: text.
Once someone's booked a meeting and opted in, a short "looking forward to Thursday — anything you want me to come ready with?" gets read in a way email just doesn't. Different rules once they're warm, though.
Rough rule I've seen hold up on first calls: ~20% go great, ~30% are positive, ~30% neutral, ~20% clearly no.
Most reps only build a follow-up path for the 20% that went great. The neutral 30% is where the quietly-winnable pipeline hides.
Try this before your next call: give Claude the prospect's role and industry, then ask "what are the 3 objections I'm most likely to hear, and the best response to each?"
Won't be perfect. But walking in having already thought it through beats improvising.
6/ None of this is exotic. The capability's been here a while. The reason it's still rare is nobody turned it into a habit. The routine is the part worth building. https://t.co/Xv4QlS6qVO
4/ Last step — have it draft 3 discovery questions you couldn't ask without the research. That's the tell. If the questions could've been asked cold, the prep didn't do anything.
5/ Ten minutes, and the rep walks in with a point of view instead of a feature list. The prospect feels it immediately — reads as "this person did their homework." Because they did.
3/ Then push it further: "What would this persona push back on, and what's the strongest response?" Now you've predicted the objection before it shows up live, instead of getting surprised by it.
2/ Feed it the company site + the prospect's LinkedIn. Ask for: 3 likely priorities tied to that person's specific role, and 2 problems worth asking about on the call. That's the whole first pass.