When Everyone's Outreach Is Written by AI

When one recruiter gets a tool that writes personalized messages at scale, that recruiter gets an edge. When every recruiter gets the same tool at about the same time, something odder happens. The personalization stops meaning anything.

That's roughly what happened to passive sourcing over the last few years. The tools promised personalized messages at scale, and technically they delivered. Each message mentioned the candidate's title, their company, maybe a LinkedIn post from last year. But the same moves showed up in everyone's outreach at once, and candidates learned to spot the pattern faster than the tools could change it. The detail that was meant to make your message stand out now marks it as automated at a glance.

It's like a concert where one person stands up to see better. It works for them, for a moment. Then the people behind stand up too, and soon the whole crowd is on its feet and nobody can see any better than before.

The numbers fit this picture. By one estimate, response rates for recruiter InMails have fallen to around 10 to 15%, from close to 30% in 2022. Another has recruiter outreach volume nearly doubling over three years. I think the connection is direct: AI made sending messages almost free, so everyone sent more, and each message became worth less. The usual reaction to falling response rates is to send even more, which speeds up the very thing that caused the fall.

It helps to follow a single message. Most get archived unopened or dismissed in a few seconds. Some get read and noted, then never acted on, because the person isn't actively looking and nothing in the message gave them a reason to start.

That last group is the expensive one. These people may really be open to a new job, just not moved by this particular message. What usually tips someone like that is a personal introduction or something changing at work, like a reorg, a pay ceiling or a new boss. A generated message can't produce either. What you do control is whether your first contact gives them a reason to engage, and volume-first sourcing has mostly stopped trying.

The incentives point the wrong way as well. Sending lots of outreach looks like activity. The numbers that actually matter (responses, conversion into pipeline, hires per sourced candidate) get checked less often, and when they drop, the instinct is to test new subject lines rather than question the approach.

So I don't expect better messages to fix this. The difference that matters is between AI that writes messages and AI that holds conversations. A LinkedIn message competes with dozens of near-identical ones the candidate got this week, and it offers them nothing to respond to except more text.

We built Asendia because we wanted to be on the conversation side of that line, and what has always kept teams off it is capacity. A sourcer can hold only so many real conversations in a day, which is why everyone fell back on messages in the first place. Software doesn't run out of hours. When a candidate comes in through any channel, whether they answered a posting, were referred, or replied to a sourcer, Asendia calls them within hours and has a spoken conversation that adapts as it goes. It talks about their actual background against the real requirements of the role, follows up on what they say, and tells them concrete things about the job they didn't know before the call. Nobody sits in a queue waiting for a recruiter to free up, and the structured notes and call excerpts go into the ATS.

Timing is why this matters for sourcing. People who are passively open stay receptive only for a short while. If you get to a real conversation before competing outreach buries yours, and before their situation settles back into inertia, sourcing produces pipeline. If you don't, it mostly produces activity. A candidate who has actually talked through their work with you has also invested something in the process, which nobody does for a message they skimmed.

There's more on what this means at the business level in the post on the recruiting agency model splitting in two, and the piece on agentic recruiting explains why AI that carries out pipeline steps is a different kind of product from AI that helps write outreach.

I doubt passive response rates will recover through better AI messages. The tools that saturated the channel can only make the remaining receptive candidates go numb faster. The teams gaining ground, as far as I can see, are the ones who worked out that a real conversation at the right moment converts well enough to make sourcing worth its cost again. It's an operational change more than a strategic one, and it's easy to put off. Teams that do put it off will likely find their sourcing funnel producing less qualified pipeline per dollar each year, and it may take them a while to work out why.

Ready to transform your hiring strategy? Schedule a Demo with our founders today!