Why more activity won't fix your pipeline
AI can send unlimited outreach and buyers filter harder than ever. What steady, high-quality pipeline generation looks like for a startup now.
First published on The Sales Manager blog in February 2015, and updated for today.
More activity won't fix your pipeline, because activity is now the cheapest thing in sales: an AI tool will write a thousand personalized-looking emails for you before lunch. What fills a pipeline is a steady run of good conversations with people who fit, and AI is far more useful for finding and preparing those than for sending more.
I wrote about activity in 2015, and the first half of that post still stands. You only get out what you put in. Waiting for opportunities to turn up rarely produces much revenue, and your pipeline needs to be bigger than your target because you won't close everything. The second half was about quality. I said I'd rather a team had 15 high-quality conversations than 25 calls with nothing in them. Eleven years on, that second half has become most of the argument.
Activity still counts
I want to be careful here, because "quality over quantity" is the favorite hiding place of every salesperson who'd rather not pick up the phone. Pipeline is oxygen. If you do nothing, you sell nothing, and no clever philosophy gets you out of that.
For a founder doing the selling, or a first sales hire with no pipeline, no referrals and no reputation, that means volume. Lots of conversations, early, because that's how you build both the pipeline and the instinct for which conversations are worth having. What changes as you get better is the ratio. Each conversation is worth more, and you earn the right to have fewer of them. The danger is the rep who wants the veteran's ratio with the rookie's results.
AI made volume free, and buyers noticed
The easiest thing to do with AI sales tools is manufacture volume. A mediocre rep can now send ten times the mediocre outreach, every email personalized-looking, none of it involving a moment's thought. I get hundreds of these a month, and the tell is how often several of them mention the same post of mine in nearly identical words. Someone who'd actually read it wouldn't arrive in a batch of fifty saying the same thing.
Buyers have noticed. In a Gartner survey of 632 B2B buyers, 73% said they actively avoid suppliers who send irrelevant outreach, and 61% said they'd prefer a rep-free buying experience overall. Instantly, which sells cold email software, puts the average reply rate across its platform in 2025 at 3.43%. Even on a cold email vendor's own numbers, more than 96% of what goes out gets no reply.
The agents aren't a shortcut either. In July 2026, Gartner predicted that AI agents will outnumber sellers ten to one by 2028, and that fewer than 40% of sellers will say those agents made them more productive. As Gartner's Dan Gottlieb put it, "more agents will not automatically mean more productivity."
For a startup, the cost goes beyond wasted sends. Every irrelevant email to an account in your ideal customer profile teaches someone you'd love to sell to that your name means noise. At AI scale, that's how you burn through your best-fit market before you've had a single good conversation with it.
Count the conversations that move a decision
When you measure activity, you get activity. People optimize for whatever you count, so count emails sent and you'll get emails sent, with the quality of each one becoming somebody else's problem. Sales is a numbers game, sure. The number that counts is the business that comes through the door.
So build activity targets backwards from what works. My 2015 advice still holds: look at what your most successful people do, and pick the ones who consistently close, who aren't always the busiest. Work out how many good conversations it takes them to produce a deal and set the benchmark from there, instead of borrowing "50 calls a day" from somebody's conference slide.
I once managed a rep who made roughly half the calls of the people around them and beat all of them, comfortably, quarter after quarter. On the dashboard they looked like they were barely trying. What the dashboard couldn't see was who they were calling and what they were asking. Every conversation was aimed.
That also answers the question every manager asks: how do you tell genuine quality from someone avoiding the phone? Outcomes. The rep doing fewer, sharper conversations is building pipeline and moving deals. The one hiding behind "quality" has fewer conversations and nothing else to show for it.
Point AI at aim
I built my own AI outbound engine, so this isn't a technophobe talking. The rule I use is the same one I use everywhere. If a task is mechanical, processing information rather than relating to a person, the machine can have it. If it involves someone's trust, or my judgment about a specific human, I keep it.
Used that way, AI is brilliant at the preparation that used to make quality expensive: researching a business in seconds, spotting which accounts look like a fit, working out the timing, surfacing the thing worth mentioning. A small team can have those fifteen well-prepared conversations a week without giving up every evening to research. Pointed the other way, it's just a bigger hose. Which one you get comes down to what you point it at, and that choice is most of what our AI for revenue teams work is about.
What steady, high-quality pipeline generation looks like now
For a startup, this is what I'd put in place:
- A short list of accounts that fit. If you can't say why an account is on the list, take it off.
- A reason to get in touch now. A new office, a new leader, a change in their market. "I saw you've opened a second office and taken on a dozen people this year" starts a very different conversation from "I'd love 15 minutes to show you our platform."
- AI on the preparation, a person on the send. Let it research the company and draft a first version. Then someone checks the facts, because these tools get things wrong with total confidence, and decides whether the message deserves to be sent at all.
- Live conversations. Phone, video, events, and in person when you can. You can't hear hesitation in an email thread, or the lift in someone's voice when you land on the thing they care about.
- Curiosity on the call. Before most first conversations I tell myself, "I don't understand how this actually works for them yet, and I want to." Then don't accept the first answer. Buyers arrive knowing your brochure. What they can't get from a chatbot is someone who understands their situation.
- A weekly rhythm. Blocked time for pipeline every week, whatever the forecast says. Pipeline built in panicked bursts gives you a feast-and-famine quarter.
- Being findable and useful. Buyers increasingly ask AI for a shortlist before they talk to anyone, and it builds that answer from what's out there about you. Write, speak and answer questions where your buyers look, and make sure your website is readable by their AI agents.
If you need someone to build and run that engine day to day, that's what a fractional VP Sales is for. I wrote a whole chapter on this in Curiosity, and it opens by agreeing with my 2015 self, which doesn't happen often.
If you're going to count something, count conversations. They're harder to fake, and much harder to automate.
More insights
-
October 27, 2026
Should you replace your SDRs with AI?
The teams furthest ahead on AI are hiring more SDRs. What that means for your outbound, and what to hand to AI first.
-
October 21, 2026
What does a fractional CRO do in the first 90 days?
A fractional CRO's first 90 days, week by week: what to diagnose, what to fix first, where AI helps, and what you should see by day 90.