The fantasy was short-lived.

In 2024, AI hype reached its peak with promises of fully autonomous SDRs. Replace your team with bots. Run outreach 24/7. Fill pipeline with zero headcount.

It sounded too good to be true—and it was.

By early 2025, most GTM leaders had learned the hard way: replacing human SDRs with AI didn’t just fall short, it cratered response rates, damaged brand reputation, and left a wake of “AI-scented spam” across every inbox in the market.

But here’s what’s actually working: AI as a co-pilot, not a replacement.

What AI Still Can’t Do

Let’s not kid ourselves—AI can do a lot.

It can do deep research on an account. It can summarize call notes. It can score accounts. It can even write phenomenal emails.

I’d like to say it can’t qualify a buyer, or notice a stakeholder going quiet, or sense when a prospect is just kicking tires. It can’t tell us which deal to kill and which one to prioritize. It can’t adapt its tone on a tense call, or read the politics inside a buying committee.

I’d like to say it can’t do those things, but it probably can, or it will be able to tomorrow.

What it can’t do well is two things:

• Connect the dots across the entire Customer Journey
• Fact check or gut check itself on important nuances

AI doesn’t know when something is weak, outdated, or just doesn’t make sense for a situation. We can and should train it to improve on these things but we still need human intervention. (As an example, I wrote this with AI and manually edited it. In the first draft, it said AI can’t tell us which deal to kill. I know it can.)

So while everyone else was asking, “How can we replace SDRs?”

The best teams were asking, “How do we make our reps better, faster, and more focused with AI in their corner?” (Just as I’m trying to produce better content faster.)

The Foundation Still Matters More Than the Tools

Before AI can accelerate our GTM motion, we need a motion worth accelerating.

The AI SDR is the perfect example. Companies rushed to pump out AI-written emails without setting the foundation:

• What’s our ICP for each product line?
• What buyer personas actually convert?
• Can we create segments for these buyers?
• What motivated them to buy our products?
• What are the main pain points we can address?
• How have we addressed them with our customers?
• What steps will AI follow to research and write messages?
• What steps will reps follow to validate/customize those messages?

If our reps are prospecting into the wrong companies, targeting the wrong personas, using one-size-fits-all generic messaging, AI won’t fix it.

It will only accelerate the problems this creates:

• Burning good prospects
• Burning through domains

This is why our Outbound Efficiency Framework and Allbound ABM Framework start with human processes.

We help teams build a repeatable system first—then layer on tech to amplify what’s working.

That’s what enables reps to work smarter.

What a Great AI Co-Pilot Setup Actually Looks Like

The best way to think of AI is like a human assistant.

Imagine you hired the smartest, hardest-working kid out of college to help your top sales rep.

What would you ask them to do and not do?

You’d ask them to:

• Save your rep time
• Help them spend more time with prospects
• Help them generate and schedule more meetings
• Help them be more prepared for each of those meetings

You wouldn’t ask them to:

• Close deals for reps
• Run sales calls for reps
• Deliver presentations for reps
• Negotiate contract terms for reps

This should be the “area of genius” for good salespeople and precisely why we hire them. We wouldn’t dream of delegating that to an inexperienced kid fresh out of college, no matter how smart they are, so doing that with AI doesn’t make sense either.

Here’s what they could do most:

• Research identified accounts and contact
• Write extremely well researched email drafts
• Help us find time to schedule meetings and calls
• Analyze call/meeting notes and summarize them
• Help the rep keep the CRM and next steps up to date
• Help prepare drafts for discovery, presentations, etc.
• Help qualify deals reps are working to save them time

All of these things are possible if we set the foundation, give them the right accounts and contacts, and clear instructions on how to complete each step. The same is true for AI.

And it allows our salespeople to spend more time, with better prospects, AND be much more prepared to make that time more impactful. This should, by definition, increase everything from meetings booked to close rates to average deal sizes.

The difference isn’t the tool. It’s the structure surrounding it.

When we’ve done the hard work of refining our ICP, building segmented lists, and training reps to follow a consistent process, AI becomes a force multiplier.

When we haven’t, it just adds noise.

Why This Matters Right Now

GTM teams are being asked to do more with less. Efficiency is king. Every new headcount requires justification, and pipeline goals are still growing.

That pressure has led some teams to chase the illusion of fully automated sales.

But the smartest CROs have learned that AI isn’t a shortcut—it’s an amplifier.

We still need reps who know how to qualify, how to personalize, how to navigate complex deals. We just need to help them do it faster, better, and at scale.

With the right systems, the right process, and the right coaching, AI lets our best reps operate like elite athletes—with a team of analysts, writers, and researchers working behind the scenes.

The Bottom Line

We can’t swap reps for bots.
But we can make our reps 10x more effective with the right AI support.

The teams winning right now are the ones who’ve embraced AI as a co-pilot—and kept the human in the cockpit.

And it’s not just me saying this. This was my biggest takeaway from attending Pavilion’s CRO Summit the other day. It seemed everyone agreed. AI SDRs failed us because they lacked the foundation and the human touch, but AI co-pilots are the way forward.