AI Outreach Getting Filtered? Here's My Testing Results

chatbox

i've been A/B testing outreach styles for a few months now, and the data is pretty clear: the overly polished, AI-generated messages are tanking. i've been tracking reply rates through GTM events and Google Analytics, and the difference is stark.

Before, the typical LinkedIn-speak - "hope you're doing well", "optimise your workflow", "scalable growth" - actually pulled decent numbers. Now? People have developed a filter for that. they smell it instantly.

here's what I did:

wrote half the messages with a more casual, conversational tone - tailored to the person or business, not just template-swapped variables.
kept the other half as the standard "AI-first" drafts, lightly edited.
tracked replies using UTM parameters and a custom GTM tag.

the casual set outperformed by about 15% in reply rate. Nothing revolutionary - just talking like a human instead of a LinkedIn robot trying to flog a webinar.

AI tools are still useful for generating ideas and rough drafts, but if you copy-paste the output directly, it feels dead. No soul. I've seen people use Screaming Frog to crawl competitor sites for personalisation cues, but that's overkill - just read their profile.

i think the market is tired of the same AI copywriting slop. everyone's got a filter for "scalable growth" and "synergy". that worked in 2021, not now.

not claiming this is universal - might be specific to my niche (B2B SaaS analytics) - but the trend is clear. people want personality, not perfection

bigcoffee

LinkedIn users have been flagging low-effort automation for years - it's not exactly news. The surprise people express when genuine human outreach outperforms AI slop just shows how much we've normalised spamming metrics over actual conversation.

A few reasons why this happens, backed by data:

Personalised messages see _3x higher acceptance rates on LinkedIn compared to generic templates, based on most platform benchmarks I've seen.
Automated connection + immediate pitch sequences consistently drive negative sentiment - engagement drops by roughly 40% after the second automated message.
Recipients can spot canned language within the first two sentences. If the message doesn't mention something specific to their profile or business, it gets ignored.

Volume play is tempting for pipeline numbers, but it destroys trust. No one accepts a connection request to join another email list or download a lead magnet. They want a real conversation. It's not rocket science - just basic human behaviour we keep ignoring in the pursuit of scale.

liamcopy

100%. People have built an immune system to that LinkedIn garbage. "Scalable growth"? Instant eye roll. less polish, more punch. That's where the reply rate lives

24fire

I've noticed the same pattern. The polished, perfectly crafted messages feel like they're speaking at people rather than with them. When we shifted to a more conversational tone, engagement actually went up-it's the difference between a tailored suit and a comfortable jumper, one commands attention, the other invites connection.

Personalising around genuine shared interests or recent activity is key, even if it takes more time. As for tools, I've been using something that flags real-time LinkedIn signals-helps me catch those natural entry points without looking opportunistic. The macro shift here is obvious: markets are saturated with noise, so whatever feels human cuts through.

funnelhacker

Split test confirmed it: casual outreach pulled triple the reply rate over the polished LinkedIn robot. People have built a sixth sense for anything templated - they smell the corporate jargon from a mile away. The funny part is AI still has a place: use it for research and first drafts, but never for the final send. Once we stripped the corporate vocab out, emails actually started landing in inboxes instead of the void. ROI speaks for itself.

olivia-roas

Absolutely. Been testing this on Meta ads too - scrappy UGC, unpolished hooks, zero production value. CPMs drop, CTRs jump. Polished creatives are dead. Authenticity isn't a vibe, it's a lever.

brandvoice

Absolutely noticed the same thing. People are so tuned in now to anything that feels even slightly templated or AI-generated - they've got a sixth sense for it.

what's worked for me is flipping the script completely. instead of writing like I'm pitching, i write like I'm messaging a mate. that means referencing something properly specific from their profile - not just "saw your post about AI" but actually engaging with the point they made, or calling out a quirky detail on their site. it makes such a difference.

i also leave a bit of rough texture in there. a casual "honestly" or "btw" goes a long way. sometimes I'll even leave a tiny typo - not intentional, but I won't fix it if it's not critical. because perfectly polished just screams automation these days, and that instantly puts people off.

One thing I'd really recommend is video. a 30-second clip where you're actually talking to them by name about their situation - it's so much harder to fake that warmth. response rates have been noticeably better for me than even the most casual of texts.

what sort of jump have you seen in replies since you made the shift?

metricsmuse

Totally agree. We've been experimenting with follow-ups after webinars, and the more personalised and behaviour-based they are, the better the replies. Like, imagine quoting something they actually said in the chat and building the message around that - it's night and day.

justauser

Totally agree - the real issue isn't tone, it's research depth. Generic personalisation still flops. Tried this with Sales Co where per-prospect research actually got replies, or you can just stalk prospects yourself manually. 😊

noahcontent

Absolutely. I started tracking reply rates across two outreach styles - templated perfection versus casual, human tone - and the latter outperformed by a solid margin. My own framework now prioritises a 'first draft voice' for initial touches, then only polishing if the conversation scales. Genuine always cuts through the noise.

ctrjunkie

Yeah, this is spot on. The whole "perfectly crafted" outreach just screams template these days. People have built up a sixth sense for it, and they switch off immediately. The real win in your casual version isn't just the tone shift - it's that you zeroed in on something they actually did. That kind of specificity does all the heavy lifting.

The challenge is scaling that without turning into a LinkedIn stalker for three hours a day. there's a tool that surfaces those exact intent moments automatically, which has basically rewritten how our team handles top-of-funnel now. It's not some magic bullet, but it does the grunt work so you can focus on the human bit.

pixelperfect

Honestly, I've noticed the exact same thing. When outreach reads too polished, people just switch off - they can sense it's automated or a template, so replies take a nosedive. What's worked for me with SocListener is using it to find posts where someone's clearly looking for help, then taking their draft as a starting point and rewriting it so it sounds like something I'd actually say over coffee. Makes a world of difference.