Best of LinkedIn: AI in B2B Marketing CW 35/ 36

00:00:00: This episode is provided by Thomas Allgeier and Freeness, based on the most relevant LinkedIn posts about AI in BWB marketing in calendar weeks thirty-five and thirty six.

00:00:09: Freenes is a BWb market research company that supports enterprise marketing teams in unlocking full potential of their customer data with the help of

00:00:17: AI.

00:00:18: Absolutely, and it's a great foundation for what we're getting into today.

00:00:21: Yeah totally so.

00:00:22: welcome everyone to.

00:00:23: Today is deep dive.

00:00:25: We are uh?

00:00:26: We're doing something really interesting today.

00:00:28: We're looking at how AI Is just completely rewiring the B-to-B marketing landscape right

00:00:34: from top to bottom.

00:00:34: Exactly help buyers research How we do outbound how you actually structure your marketing team.

00:00:40: And our mission today is really to synthesize all this real time data From practitioners who've got founders marketing leaders people sharing actual hard lessons on LinkedIn from the front lines.

00:00:50: Which

00:00:50: is the best kind of data, honestly because you know we're not talking theory here.

00:00:54: these are curated On The Ground insights.

00:00:57: We've got real numbers Real failures and actual wins From the market right

00:01:00: now.

00:01:01: Yeah so let's just jump right in Because were basically looking at this concept Of the invisible funnel And have to start At top that funnel Where buyers Are actually being sold too.

00:01:10: The traffic shift.

00:01:11: Yes Thomas Seifert, he's the CFO of Cloudflare shared this staggering statistic.

00:01:17: He pointed out that non-human traffic on their network actually surpassed human traffic in Q two of twenty twenty six.

00:01:24: Surpass it just

00:01:25: wow right.

00:01:27: and he warns that machine traffic could run like a thousand times higher than human traffic In just five years.

00:01:33: So my first reaction to this was wait if the internet is just bots talking to bots are we as marketers basically?

00:01:39: Just Throwing billboards up in the matrix like

00:01:41: screaming into an empty room.

00:01:43: Yes,

00:01:43: are we marketing to an empty?

00:01:44: Room?

00:01:45: I mean it's a fair question.

00:01:47: But the reality is actually the total opposite because commercially those bots represent really highly qualified human intent.

00:01:55: Lonnie Stark over at Adobe shared some wild data on this.

00:01:59: she noted that AI referred retail traffic was up sixty two percent year-over-year In July.

00:02:04: okay, sixty Two percent

00:02:05: yeah.

00:02:05: but here's The Kicker That AI-referred traffic is converting sixty percent higher.

00:02:11: Wait, really?

00:02:12: Sixty percent higher conversion.

00:02:14: Yep

00:02:14: and it's generating fifty three percent more revenue per visit.

00:02:19: The Adobe data essentially shows these visitors are four point four times more likely to actually take action.

00:02:24: That is I mean that implies the AI Is basically doing the bottom of the funnel qualification for us before the human ever even clicks

00:02:32: over precisely because think about how you use an AI agent.

00:02:35: You don't just say give me a list of software, right?

00:02:38: You give it your budget your constraints Your tech stack.

00:02:41: so by the time the bot sends a click to your site that Human has already sold great

00:02:45: which honestly completely breaks traditional SEO.

00:02:48: It

00:02:48: totally shatters

00:02:49: it because if you look at your organic traffic and its flatlining, You might think oh I have an SEO problem but actually a AI visibility problem or generative engine optimization.

00:03:00: Geo?

00:03:01: Yeah Marja Voje at RS looked into this.

00:03:03: she found that twenty eight point three percent of the pages chat GPT sites.

00:03:07: most they literally zero organic visibility on Google Zero!

00:03:12: That is wild

00:03:13: Nothing, it's completely disconnected from traditional search rankings.

00:03:17: And she actually found that YouTube mentions predict AI visibility better than any other signal

00:03:23: Which is so funny because you know most b to be teams treat youtube as a total afterthought like dumping ground for webinars.

00:03:31: right just uploading boring zoom call once a quarter.

00:03:35: But why are the AI models so obsessed with video transcripts?

00:03:39: Well, it's about data density.

00:03:40: right A transcript of real conversation is nuanced.

00:03:43: It has expert vocabulary not just some thin keyword stuff.

00:03:47: blog post LLMs just crave that context.

00:03:50: heavy dense text.

00:03:52: That makes a ton of sense, and it actually ties into who is creating the content too.

00:03:56: Tarana Heath shared some LinkedIn data showing that seventy-five percent of LinkedIn citations in AI searches, seventy five percent come from individual member profiles not official company pages.

00:04:07: I love that step

00:04:07: And thirty-five per cent of those comes from users with under ten thousand followers.

00:04:11: Because the AI doesn't care about your corporate brand guidelines.

00:04:14: It's looking for real expertise.

00:04:16: If you're lead engineer breaking down complex problem The AI scrapes because it's authentic.

00:04:22: It's not a sanitized press

00:04:23: release.".

00:04:24: So

00:04:24: relevance really does beat brand size now.

00:04:27: Annie York actually had a great concrete story about this, she worked with the really small parent provider Terrible SEO.

00:04:34: they could never beat The Giants on Google for like best payment provider.

00:04:38: so what do

00:04:39: They got?

00:04:39: super specific best EC terminal for pharmacies.

00:04:45: Oh, nice!

00:04:45: Yeah and just getting that specific recommendation in chat GPT it resulted in twelve solid leads.

00:05:02: GEO isn't just about tweaking keywords in an article, it's a bland authority problem.

00:05:06: Yeah if the AI engines don't have reason to trust you... Just pumping out answer first.

00:05:11: content isn't going make you visible?

00:05:12: Exactly!

00:05:13: It is like trying to win a Michelin star.

00:05:15: The food critic is the A.I.. It doesn't care how big your neon sign outside

00:05:19: Right.

00:05:20: So okay If inbound research is shifting to these answer engines, we have look at outbound too.

00:05:25: Because how sellers are reaching out is shifting just as fast.

00:05:28: Oh man the outbound shift is crazy!

00:05:30: It really is... The whole narrative around full autonomous AI-SDR.

00:05:36: that pitch isn't a full blown retreat right now

00:05:39: Thank goodness.

00:05:40: honestly

00:05:41: Nick Abraham pointed this out There's this company Artisan and they spent like Two million dollars heavily promoting this massive slogan stop hiring humans.

00:05:51: Oh, I

00:05:51: remember seeing those ads everywhere.

00:05:53: Right

00:05:53: but recently they quietly retired the slogan and they actually hired their first human BDR.

00:05:59: Wow you can't make that up

00:06:00: You really?

00:06:01: Can't?

00:06:01: yeah ramp too.

00:06:02: They shut down there internal AI SDR program which they called OA because it just couldn't handle The complexity of an actual sales motion.

00:06:09: so

00:06:09: the dream of Just pressing a button then having a bot print pipeline is

00:06:13: dad pretty much yeah.

00:06:15: vignesh w shared some numbers on this.

00:06:17: he noted that while forty one percent of enterprise teams tried adopting aist rs only twenty two percent fully replaced humans.

00:06:24: so the vast majority just rolled it back.

00:06:26: yet they reverted to hybrid models because the market is brutal right now.

00:06:31: cold outbound reply rates were like, eight to twelve percent back in twenty-twenty six.

00:06:36: They've dropped it two or four percent.

00:06:38: That is just a massive collapse.

00:06:41: But wait why did the autonomous dream fail so badly?

00:06:43: Was AI writing robotic weird emails?

00:06:47: See that's what everyone assumes.

00:06:49: but the AI wasn't really the problem.

00:06:51: The underlying data was the problems.

00:06:53: Oh of course

00:06:54: Kyle Ace shared this story about Launch Darkly.

00:06:57: they tried rolling out an AISDR and had to scrap the whole thing.

00:07:00: And the root cause wasn't the language model hallucinating, it was duplicate accounts in their CRM.

00:07:06: Oh no!

00:07:06: I can see exactly where this is going.

00:07:08: Yep

00:07:08: The AI couldn't tell difference between a new prospect and an existing customer.

00:07:12: so just firing off aggressive cold emails to people who were already paying for product.

00:07:17: That's a nightmare.

00:07:18: If your CRM has held together by duct tape An autonomous bot will automate data hygiene failures at scale.

00:07:24: It really shows that fundamentals still matter.

00:07:27: George Custeria posted an enrichment benchmark, which perfectly illustrates this.

00:07:31: He tested Prospero and it scored a ninety-eight point eight with one hundred percent deliverability And only costs like five dollars per thousand emails

00:07:41: which is incredible.

00:07:41: Right,

00:07:42: so it proves that good data beats a good prompt every single day of the week?

00:07:46: Absolutely!

00:07:46: Oh and by the way real quick if you're enjoying these insights and want to keep up with how the B-to-B landscape is shifting.

00:07:52: make sure to hit subscribe.

00:07:53: So don't miss our future deep dives.

00:07:55: Yes definitely subscribe because this stuff is changing weekly And actually That brings up the hybrid formula

00:08:01: right The fix for the autonomous failure.

00:08:04: Enzo Carasso shared how his agency booked twelve meetings in a single day for a client.

00:08:09: Twelve

00:08:09: Meetings

00:08:11: In One Day?

00:08:12: By pairing an AI, SDR directly with the human SDR The takeaway there is that the AI creates volume.

00:08:21: It takes all shots on goal But the human is here to take important shots.

00:08:25: Human catches CRM dupes before email goes out.

00:08:29: So that leads us right into the third big theme, which is how you actually rewire your marketing org for this.

00:08:36: Because if the winning strategy is humans and AI working together... How do you manage that?

00:08:43: Structurally I mean!

00:08:44: It's

00:08:44: a huge bottleneck.

00:08:45: Caroline Healy brought this up with some BCG data.

00:08:48: Ninety-six percent of CMOs say AIs transforming their function but only eight percent are running Multi-agent autonomous campaigns eight

00:08:58: percent.

00:08:58: That is a massive gap.

00:09:00: It's

00:09:00: huge and Healy argues that it because you have to assign a named human owner To every single agent.

00:09:06: Oh, I like not just a prompter but someone who was officially accountable for what?

00:09:11: Accountability?

00:09:12: yeah Because otherwise you just get like assistant sprawl

00:09:14: Yeah spots

00:09:15: everywhere doing random things

00:09:17: which is exactly What Justin Norris is fighting against.

00:09:19: he talked about moving away from that messy Assistant centric model into a really clean Four layers skills architecture.

00:09:26: Okay, four layers break that down for

00:09:28: me.

00:09:28: So layer one is connectors basically the API hooks.

00:09:31: so The bot's ability to actually read the CRM or send an email right.

00:09:36: then later.

00:09:36: two his platform skills.

00:09:37: This is the basic business logic like never e-mail a current customer

00:09:42: ah?

00:09:42: They'll launch strictly fix

00:09:44: exactly.

00:09:44: Layer three has worked.

00:09:45: flow skills this is the actual job Like draft and outreach email based on LinkedIn post.

00:09:51: okay got it in the fourth

00:09:53: context skills.

00:09:54: This is your brand messaging, you're tone of voice.

00:09:57: Your specific product detail... That's

00:09:59: so smart!

00:10:00: Because if you update your messaging You don't have to go retrain twenty different agents.

00:10:03: So just update the context layer

00:10:05: Precisely.

00:10:06: It keeps everything centralized.

00:10:08: But wait when building out that complex Of a stack What AI model do even use?

00:10:13: There's new one dropping every Tuesday.

00:10:15: Well

00:10:15: Orin Greenberg ran massive hundred and forty four question Go-to market benchmark To figure it

00:10:21: all.

00:10:21: Oh wow what did he find?

00:10:23: He found that Claude Opus is only one point four percent more accurate than Claude Sonnet for analytical GTM tasks.

00:10:30: One

00:10:30: point four per cent.

00:10:30: but wait, opus is way more expensive right?

00:10:32: It

00:10:32: costs two and a half times more!

00:10:34: That's insane.

00:10:36: so you're paying massive premiums for tiny bump in accuracy

00:10:39: Right...and don't even really need the AI to be perfect because human oversight is what actually multiplies value.

00:10:46: Yes..the Human In The Loop Neil Patel actually ran a massive test on this.

00:10:51: They looked at seven hundred and forty-four articles across sixty eight websites.

00:10:55: That's

00:10:55: huge sample size

00:10:56: Huge, in the result Taking just twenty minutes to have a human edit, an AI generated draft produced four hundred and forty-four percent more traffic than publishing the raw AI draft.

00:11:08: Four Hundred and Forty-Four percent?

00:11:10: Right it sounds like AIs basically The Ultimate Sous Chef.

00:11:12: It can you know chop vegetables prep ingredients at lightning speed but You still desperately need a human head chef To actually taste food And plate dish before goes to customer.

00:11:22: That is perfect analogy.

00:11:23: The AI does the prep, the human ensures the quality and this whole structural shift in how teams operate is causing massive ripple effects on the broader market.

00:11:32: The

00:11:32: talent migration?

00:11:33: Oh yeah!

00:11:33: And the tool consolidation.

00:11:35: Adam Schoenfeld tracks this stuff... ...and his August tracker he saw that conversation around proof of humans spiked a hundred and seventy-six percent.

00:11:43: Wait,

00:11:43: Proof Of

00:11:43: Human?!

00:11:44: Yeah basically verifying that a human was involved in process.

00:11:48: This spiked right after Anthropic announced their watermarking.

00:11:52: Meanwhile The hype around signal-based prospecting, just pure automation fell by half.

00:11:57: It dropped from eight point three percent to four point one

00:12:00: percent.".

00:12:00: The market is course correcting and you can see it in where the top talent.

00:12:07: He just left a twenty-year career at Google where he managed to fifteen billion dollar ad portfolio.

00:12:13: Huge job!

00:12:13: Right, and he left to join Hightouch betting entirely on agentic marketing built directly on data warehouses

00:12:20: Because data is the foundation Exactly.

00:12:22: And Corey Stryker too.

00:12:24: He left ClickUp after seven years of hyper growth To join Qualified and work specifically on Piper which was their AISDR.

00:12:31: These guys are betting entire careers On this hybrid agente shift

00:12:34: because that's what competitive advantage now.

00:12:37: Jason M. Lemkin put it really well, he said that in B-to-B today the advantage goes to whoever has the absolute best agent and their space.

00:12:44: so

00:12:44: It's not just a commodity.

00:12:45: yet

00:12:46: Not at all everyone has access to this same baseline.

00:12:49: LLMs sure but solving A massive customer problem with an agent That is viscerally better deeply integrated into your specific workflows?

00:12:58: THAT Is how you win.

00:13:00: So its about the architecture not Just The Model

00:13:02: Exactly

00:13:03: Wow.

00:13:04: Okay, so to wrap all of this up I want to leave everyone listening with a final pretty provocative thought.

00:13:08: yeah Eric Buckley pointed out a Gartner prediction that is just mind-blowing.

00:13:12: they predict the by twenty twenty eight AI agents will intermediate ninety percent of BtoB purchases.

00:13:18: Ninety percent?

00:13:19: Ninety per cent!

00:13:20: That's fifteen trillion dollars moving through these agents.

00:13:23: So we are literally from business to business btb to business agent.

00:13:26: How do you do A?

00:13:27: Exactly...so The Final Question You Really Need To Mull over This Week Is This Is your brand-building authority that a machine will actually respect?

00:13:35: Or are you just optimizing for a human buyer who is no longer making the first cut.

00:13:40: That's the real question, because The Invisible Funnel is already here!

00:13:44: It really is... If you enjoyed this episode new episodes drop every two weeks.

00:13:49: Also check out our other editions on field marketing channel and partner marketing account based marketing MarTech go to market And social selling.

00:13:57: it has been great deep dive today.

00:14:00: Thank you so much for joining us and remember to subscribe.

00:14:03: See catch the next one!