AI SDR is not AI. It's spam with a better grammar checker.

I'm done being polite about this.

AI SDR is not AI.

It's spam with a better grammar checker.

Three vendors in the category have raised over $170M in venture funding since 2023: 11x, Artisan, and Regie.ai. In the same window, B2B cold email reply rates dropped from 7% to 3.43%. Open rates dropped from 36% to 27.7%. Gmail rewrote its bulk sender rules in February 2024. Microsoft followed in May 2025.

The category is selling more of a thing that works less.

This is a manifesto. The case against the spam cannon. The case for the seller who does the work. And the case that the math on the volume motion is not getting better at scale. It is getting worse. By design.

I'll show you what landed in my inbox in the last 30 days. Then I'll show you the numbers underneath. Then I'll plant a flag.

This is not a copy problem. It's a structural problem.

Look at these. Read them. Then ask yourself if any of this is sales.

This is not bad copy. Not a personalization gap. Not something a better prompt fixes. This is what happens when you stack GenAI on top of broken data and call it a salesperson.

The list is wrong. The context is missing. The signal is fake. The "personalization" is a scraped LinkedIn headline reworded by an LLM. The "intent" is one click on a third-party site that gets resold to 40 vendors at the same time.

Here is what the data says about that motion. Cold email reply rates dropped from 7% in 2023 to 3.43% in 2026 (Instantly's benchmark report, billions of sends analyzed). Open rates dropped from 36% to 27.7%. 17% of cold emails never reach the inbox at all. AI-generated emails, when the buyer can detect them, get 90% lower response rates than human-written ones. The buyers are catching up faster than the prompts are.

The platforms agree. In February 2024, Gmail rewrote the rules for anyone sending more than 5,000 messages a day: mandatory SPF + DKIM + DMARC, one-click unsubscribe, spam complaint rate below 0.3%. In November 2025, Google escalated from spam-folder filtering to permanent rejection at the SMTP level. Microsoft followed in May 2025. The infrastructure of email itself is now actively pushing back on the volume motion.

You can't fix that with a better prompt. You can't fix it with a "humanizer" model. You can't fix it with another follow-up sequence. The foundation is broken because the people building these tools never understood what sales was in the first place.

Sales has been automated the same way for years.

Faster lists. Bigger sequences. More automation. All of it built for the seller. None of it built for the buyer.

The buyer was never the customer. The buyer was the target.

AI SDR is the worst version of that pattern. It scaled the one-sidedness past what the channel can carry. Email is not dying because of AI. It is dying because of years of tools built by people who think buyers can be reduced to industry, job title, country, and a layer of poor GenAI copy on top.

Anyone who builds for sales like that does not love the craft.

So what is sales, actually?

Sales is research. Sales is qualification. Sales is disqualifying more accounts than you qualify.

Sales is curiosity. Judgment. Restraint. Timing. The discipline to not send the email when you have nothing useful to say.

Sales is sounding like a peer. Not a pitch bot.

Every real seller I know does the work before the outreach. They study the account. They find the angle. They earn the first 20 seconds. They pick up the phone. They walk into the meeting with a point of view that's been earned, not generated.

You can't automate that with a domain rotator and a 4o-mini wrapper.

I have never met a top performer who didn't deeply understand their customers' business, follow industry signals, build granular playbooks by persona, and take pride in solving complex problems. That's the craft. That's the part the AI SDR crowd has decided is replaceable.

It isn't. It never was.

Volume is the only motion that gets worse when you scale it.

Every other operations problem in business gets better with scale. More servers means more throughput. More warehouses means faster delivery. More reps on the phone means more calls.

Outbound email is the exception. The volume motion is the only sales mechanic in modern GTM that mathematically gets worse the more you do of it. This is not an opinion. It is a stack of structural facts.

Deliverability degrades with volume. Cold campaigns at scale routinely produce spam complaint rates of 0.5% to 1%. Gmail's bulk sender threshold is 0.3%. Google recommends staying below 0.1%. The math means a team blasting 100,000 emails a month is not slightly over the threshold. It is structurally at war with Gmail's policy. Since November 2025, the consequence is no longer the spam folder. It is permanent SMTP rejection at the front door.

Reply rates have collapsed. From 8.5% in 2019 to 7% in 2023 to 3.43% in 2026. That is not a temporary dip. That is a channel doing what every outreach channel has done before it. Email is at its AI moment the way fax was at its junk-mail moment in 1995.

The unit economics get worse, not better, with more reps. Doubling the team blasting from burner domains does not double meetings booked. It doubles the spam complaint surface area. It doubles the rate at which inbox filters learn your patterns. It doubles the speed at which fresh domains burn out and need to be rotated. The 730-emails-per-meeting ratio that the volume motion celebrates is not a fixed cost. It is a number that decays with every additional send, on every additional domain, into every additional inbox.

AI-generated emails are the accelerant, not the fix. When buyers can detect AI, those emails see 90% lower response rates than human-written ones. The "humanizer" wrapper is an arms race the buyer is winning. Every new prompt template gets pattern-matched into the same spam pile inside 30 days.

This is why "just add more SDRs to the AI SDR" is not a strategy. It is the same hand on the same lever, pulling harder. The volume motion is fragile to scale. The augmented motion is anti-fragile to it. Cody and Magnus, on the phone, with agents doing the prep, do not see their numbers degrade when they make more calls. The mechanic compounds in the other direction. Better research means better calls means more meetings means better research on the next round.

Scale the right motion and the numbers compound. Scale the wrong one and they decay. The category sold the second one as the first.

Top of funnel is the entire game.

The list. The signal. The opener. Get those right and the rest of the funnel takes care of itself. Get them wrong and no follow-up sequence, no nurture flow, no "humanizer" prompt saves you.

And the game at the top of funnel is not volume. It is laser-focused, deeply personalized outreach to the small set of accounts where the timing is right and the fit is real.

The math is not subtle. Manually researched micro-campaigns produce reply rates above 20%. Volume campaigns celebrate 3%. The volume crowd will tell you they make it up on scale. They won't tell you that the other 97% are now actively trained to ignore the sender. That damage is permanent.

Volume optimizes for the bottom 97%. Focus optimizes for the top 3%. Pick.

Email and LinkedIn are suffocated. The phone is wide open.

Email and LinkedIn are not technically dead. They are suffocated. Bots flooded them. Gmail and Microsoft fought back with strict deliverability policies. Buyers got fluent at recognizing AI-generated outreach in about two emails. The channels still exist. The volume game on them is over.

The phone is the new blue ocean. Less noise. Direct conversations. Buyers actually pick up when the opener is good.

There is a generational catch. Half the reps entering sales today don't even have the call button visible on their phone. They didn't grow up with the phone as a conversation tool. They need training. The phone is wide open and nobody is calling. That is the entire opportunity.

And this is where the augmented motion actually shines.

AI agents do not write the call for the rep. They prepare it. Account Research finds the angle. Contact Research finds the human behind the title. Play Copywriting generates a research-based cold call opener calibrated to that persona. The rep walks in sounding like a peer, not a script reader.

The augmented motion

Five agents. One sharper rep.

Each agent passes deeper context to the next. The rep makes the call.

Account Qualification Filters in or out before a single touch

→

Account Research Finds the angle, the signal, the why now

→

Contact Finder Persona-fit contacts, not random titles

→

Contact Research Finds the human behind the title

→

Play Copywriting Research-based opener, persona-calibrated

→

Cold call opener Grounded in real research, ready in the rep's pocket

That is working right now. That is where the meetings are coming from. The full evolution of why this is happening, channel by channel from fax to AI spam, is laid out in Sales Always Kills Channels by Automating Them to Death.

The buyers on the other side aren't dumb.

They're founders. They're VPs of Sales. They're CMOs and CFOs and Heads of RevOps. They get 15 cold emails a week, on average. They've seen a thousand of these. They know the burner domain trick. They know the "I saw your post" cold open is fake. They know the "quick question" subject line is a sequence.

Here is what nobody in the AI SDR category likes to say out loud: those buyers are paying for this motion. Not the senders. The buyers.

At 15 cold emails per week, at 10 seconds each to read the subject, scan the first line, and delete, every B2B decision-maker spends roughly two hours per year clearing AI SDR junk out of their inbox. The largest survey of cold email recipients found that 71% of those emails fail on relevance. Multiply two hours by the ten million B2B decision-makers in North America and Europe. That is a measurable productivity tax extracted from the economy to harvest a few replies.

Worse than the time is the decision overhead. Every fake "I noticed your recent post" forces a real-or-not evaluation before the buyer can delete it. Every "humanizer" wrapper trains the buyer to read the next email more defensively. The cumulative cost is that buyers stop trusting the channel at all. That is why open rates dropped from 36% in 2023 to 27.7% in 2026. Buyers learned. The channel paid the price.

Read that math the other way. 730 emails to book one meeting means 729 people received an email they did not want and did not ask for. The collateral damage is not a side effect of the volume motion. It is the model. The 729 are not a rounding error. They are the design.

And then the rep who does it right gets punished for it. The seller who spent two hours researching the account, building a real angle, writing one email to one person. That email lands in the same junk folder as the 100,000 the AI SDR blasted that morning. The buyer cannot tell the difference. The inbox filter cannot tell the difference. The legitimate seller pays the externality of the volume motion.

That is the damage. The few positive replies are not worth what they cost everyone else to harvest.

Evergrowth is the Anti-AI SDR.

We are not a faster spam cannon. We are not a cheaper Apollo. We are not a Clay alternative. We are not a "personalization layer." We are not an AI SDR with a better prompt.

We are a workspace where 13 specialized AI agents act as digital colleagues for the sales rep. They handle the research, the qualification, the signal monitoring, the contact discovery, the call prep, the playbook generation. They give the rep back the 6+ hours per week that should have been spent selling in the first place.

The rep is still in the seat. The rep still makes the call. The rep still closes the deal. The rep is just sharper, faster, and dangerous because they walked in with context the other side didn't expect.

AI SDR replaces a job that doesn't exist. Evergrowth elevates the one that does.

If you sell AI SDR: come sell to me like a human. I'll take the call.

If you buy AI SDR: ask your reps what they think. If they're not in the room when you sign the contract, you're already lost.

If you build for the craft: we're hiring, we're partnering, we're listening.