AI SDRs after the hype: what’s actually working
The first wave of AI SDRs promised a pipeline machine that never sleeps. Plenty of teams got burnt domains and annoyed prospects instead. Here’s what separates the AI that works from the AI that doesn’t.
What you will learn
- Why did so many early AI SDR rollouts disappoint?
- Does adding more AI to a sales team keep adding more productivity?
- Which sales development jobs is AI genuinely good at today?
In this article
Two years ago, the pitch was irresistible. Plug in an AI SDR, give it your ideal customer profile, and wake up to a calendar full of meetings. No hiring, no ramp, no churn.
Plenty of teams tried it. Many quietly switched it off.
That doesn’t mean AI in sales development has failed. It means the first wave pointed it at the wrong job.
What went wrong with the first wave
Most early AI SDRs were built to do one thing: send more outbound, faster. That turned out to be the part of sales development that least needed speeding up.
- Volume isn’t the bottleneck. Inboxes were already full of generic cold email. Doubling the volume mostly doubled the noise.
- Mailbox providers fought back. Google, Yahoo and Microsoft now enforce strict rules on bulk senders, and high-volume AI sequences are exactly what those rules catch. More on that in our post on cold email.
- Personalisation at scale started to look the same. When every email opens with a line about the prospect’s latest LinkedIn post, none of them feel personal.
- Meetings booked isn’t revenue. Some tools hit their meeting targets with prospects who were never going to buy, which simply moved the waste from SDRs to account executives.
More AI doesn’t mean more productivity
Gartner put a number on the problem. It predicts that by 2028 AI agents will outnumber sellers ten to one, yet fewer than 40% of sellers will say those agents made them more productive.
Its analysts describe a ceiling: past a certain point, stacking more bots and prompts onto a workflow overwhelms the people using it. Their advice is to pilot narrow use cases with clear goals, fix your data first and judge success on outcomes, not activity.
The question isn’t “how much AI?” It’s “which job, done how well?”
Where AI is genuinely earning its keep
The teams getting results in 2026 tend to give AI narrow, well-defined jobs where speed and consistency matter more than volume:
- Responding to inbound instantly. A buyer who has just asked for a demo is the warmest lead you’ll get. AI can answer in seconds, at any hour.
- Running a consistent first conversation. Same discovery questions, same honest answers, every time, whether it’s 9am Monday or 10pm Sunday.
- Research and preparation. Summarising an account, a call or a thread so a human walks in prepared.
- Clean handovers. Capturing what the buyer said, what they need and what they asked, so the next call starts where the last one ended.
Notice the pattern. These are jobs where the buyer has already raised their hand, or where AI supports a human rather than replacing one.
Questions to ask before you buy an AI SDR
- Is it pointed at inbound or outbound? Answering people who asked to talk to you is a very different risk from emailing people who didn’t.
- How does it hand over to a person? If the answer is vague, your reps will spend their time cleaning up.
- What happens when it doesn’t know? Good AI says so and books a human. Bad AI improvises.
- How will you measure it? Pipeline and qualified second calls, not emails sent or meetings booked.
The bottom line
The first wave of AI SDRs tried to automate the salesperson. The second wave is more modest and more useful: it automates the moments where speed matters most and gives people better information for the conversations that need a human.