What is actually working?
A few years ago I built what I thought was a smart sales pipeline to reach out to destination marketers for my travel platform Map Your City. Custom sequences in Mailchimp, automated follow-ups, the works. It ran beautifully. And at some point I realized I had absolutely no idea what was happening inside it. Who was where. What was working. The automation had quietly taken me out of my own sales process.
AI in sales is not about "do more outreach, faster"
AI in sales is being sold as "do more outreach, faster." But more outreach isn't the goal. More right conversations are.
Most AI sales tools optimize for volume and the appearance of personalization ("I saw you spoke at X event!" or "I noticed you replied to X post"). Prospects have learned to recognize it. Worse, you stop knowing your pipeline.
To be clear, AI is genuinely transforming how sales and go-to-market teams operate. An AI agent that qualifies incoming leads at 2am, follows up consistently, and never has an off day doesn't need to be "better than a human" to deliver serious value.
AI is also getting good at the learning loop — figuring out what messaging resonates, with which type of prospect, at which stage of the conversation. That's exactly the kind of pattern recognition that humans do slowly, inconsistently, and usually only after losing a few deals.
Where it breaks down is one specific thing: putting the whole thing on autopilot. When AI is prospecting, writing, and sending without you in the loop, relevance goes out the window — and so does your deliverability. That's not AI-assisted outreach. That's just outsourced spam.
"You sent one email to 1,000 people. That's not personalization at scale just because you mail merged their name into it." — Sam Jacobs, founder of Pavilion
He's right. And it's gotten worse since AI made sending 10,000 emails almost free.
I now use Claude + Notion + a spreadsheet. Smaller subset of prospects. More deliberate. Better conversion. And I actually enjoy it.
So the question isn't "should I use AI for sales?" It's "where is AI helping me be more relevant and where is it letting me be lazy about it?"
The tool
My own stack:
- Claude + a scraper tool: a simple scraper that works great to grab local info and emails from Google Maps or any social platform. I feed this to Claude to prioritize and add it to a spreadsheet.
- Spreadsheet: overview, prioritization, nothing fancy. But Claude nowadays is integrated remarkably well with Microsoft's tools. I always research a prospect before outreach, drafting a message that's actually specific for that person, sense-checking my approach.
- Claude Outreach Skill: generates a date-organized brief of outreach tasks and follow-ups for my prospect list and updates it.
- Notion: tracking where each prospect is, what I have sent, what I am waiting on.
Why it works: I stay inside the process. I know my pipeline. I make the calls.
The move
Pull up your last 10 outreach messages — email, LinkedIn, whatever.
Ask yourself one question: "Does this say something actually relevant to this person or this audience, right now?"
If yes → great, AI is helping you be more relevant. Whether you're reaching out to one specific prospect or a well-defined niche with something that genuinely fits their world, it's working.
If no → AI is probably doing the writing, not the thinking. Flip it: use AI to research and prepare first, then write — or brief the AI properly on what makes this audience tick before it drafts anything. You can build variations for different segments or situations, run them by AI for feedback. More than once it's stopped me sending something with a typo too.
Bonus move: if you use an automated sequence, go back in and ask: "does every message in here say something worth reading, or is it just filling a cadence slot?"
From the Field: don't manage a system — start managing relationships
I stripped it back after Map Your City. Now it's Claude, Notion, and (yes) a spreadsheet. Fewer prospects in play at any one time. Every message I send, I know exactly why I'm sending it. And I get better conversion. More importantly, I actually look forward to doing it.
The tools didn't change the outcome. Staying inside the process did.
If you're still thinking that letting AI handle the outreach while you step back will bring in leads at near-zero cost — I hate to break it to you. There's also a mechanical cost that most people don't factor in. Your domain reputation is built on engagement: opens, replies, people actually wanting to hear from you. Send enough ignored or marked-as-spam emails and you start losing deliverability across the board. The AI-generated outreach that "saves time" today is quietly making tomorrow's emails harder to land. Even the good ones.
Implementing AI in your sales process is as much a company culture challenge as it is a technical one. Most companies go straight to "which tool?" without asking whether AI is even the right solution category for that part of the process. Want to see if yours is doing the same?