Why smart CROs aren't fully automating their SDR teams

The short answer: the agents can do the work, and several CROs I've spoken with privately are still choosing to keep humans in those seats. The SDR team is their talent pipeline. It's where the next generation of AEs learns the craft. Automate the whole thing and you save money for two years, then discover you closed the classroom where sellers are made. The fix isn't avoiding AI. It's deciding what the junior seat is for.

Here's the longer version.

The conversations happening in private

Over the past few months I've had the same conversation with several CROs. Each one could automate their SDR function further, maybe fully. They've tested the agents. The meetings get booked. The math works on paper.

And each one is choosing not to.

The reason has nothing to do with technology. The SDR team is where their next AEs come from. It's where people learn to hear a brush-off and keep going, to earn thirty seconds with a stranger, to feel the difference between interest and politeness. Take those seats away completely and in two years there's nobody coming up who learned the job from the ground up.

Nobody says this on stage, because "we kept humans for the talent pipeline" doesn't sound as bold as "we replaced the team with agents." But it might be the smarter long-term call.

The engineers got there first

Software is running this experiment ahead of us. The code writes itself now, and some junior engineers ship fast without ever learning why the code works. Senior engineers have started asking where the next senior engineers will come from if juniors never wrestle with the basics.

Sales should watch that closely, because our version is worse. Code at least fails loudly. A rep with weak judgment fails quietly, one mishandled objection and one misread room at a time, and it shows up two quarters later as a soft pipeline nobody can explain.

I run an AI sales company, so you'd expect me to wave this worry away. I can't. The risk is real.

Where judgment actually comes from

Judgment in sales is built in live moments. The pushback you didn't expect. The champion who goes cold. The pricing conversation that turns. The moment you realise the person nodding along has no power to buy.

No script covers those moments. You get good by being in them, over and over, and by someone helping you understand what happened afterward. That second part matters as much as the first. An at-bat you never review teaches you almost nothing. Most reps have years of experience and only a few hours of anyone actually working through their meetings with them.

One more thing is true, and it matters for how you use AI: nobody ever learned to read a room by updating the CRM at 9pm. Admin never built a seller. Live conversations did.

The middle path

Put those pieces together and the answer stops being "automate or don't."

Let agents carry the work that never taught anyone anything: the prep documents, the follow-ups, the CRM updates, the chasing. That buys your people more time in live conversations, which is where judgment grows.

Then coach every one of those conversations, while they're fresh. That's the half most teams skip. More at-bats without review just means more unexamined habits. A junior seat becomes a classroom when someone works through the tape.

So the middle path looks like this: agents carry the work, humans hold the conversations, and coaching turns the conversations into skill. The seat gets more productive and it still produces AEs. You don't have to choose between this quarter's cost line and 2028's AE bench.

This is the design behind how we built Grw: it does the operational work around every deal, and every rep gets coached on every real call, against the same standard. The point of removing the busywork was never fewer sellers. It was better ones, faster.

Questions to ask before you automate a junior seat

If you're weighing how far to take AI in your SDR or junior AE function, four questions help:

• What is this seat for? If it exists purely to book meetings, agents will win on cost. If it also exists to grow your future AEs, price that in before you cut it.
• Where will your next AEs come from? If the answer is "we'll hire experienced ones," remember everyone else is planning the same thing, from a shrinking pool that nobody is training.
• What work in the seat builds skill, and what work just fills the day? Automate the second list aggressively. Protect the first.
• Who reviews the live moments? If nobody coaches the calls, the seat isn't a classroom anyway, and that's the thing to fix first.

The CROs quietly protecting their SDR teams aren't behind on AI. They're seeing something the org chart doesn't show: the place where sellers are made. Keep the classroom open. Just stop making the students do the paperwork.