Best of LinkedIn: AI in B2B Marketing CW 19/ 20

00:00:00: This episode is provided by Thomas Algeier and Frennis, based on the most relevant link-in posts about AI in B to D marketing.

00:00:07: In calendar weeks, nineteen and twenty.

00:00:09: Frennis is a BDB market research company that supports enterprise marketing teams in unlocking full potential of their customer data with help from AI.

00:00:18: You can find more info in description.

00:00:20: Alright let's get right into it.

00:00:21: We've got lot ground cover today.

00:00:25: So welcome to the deep dive, everyone.

00:00:27: Whether you are a seasoned CMO or an emerging GTM engineer today's your playbook.

00:00:33: we're distilling The top AI in B-to-B marketing trends that have basically taken over LinkedIn these past two weeks.

00:00:40: Yeah We're cutting through all of that vendor fluff.

00:00:43: yeah want to map out exactly how AI is rewiring the btb Marketing space right now.

00:00:47: and I mean were looking at three core themes today.

00:00:49: We've got the evolution of AI visibility, so that shift from SEO to AEO.

00:00:54: Then the whole reality check hitting AI SDRs and outbound motions.

00:00:58: And finally The rise of GTM agents.

00:01:00: and well moving to a system-centric architecture.

00:01:03: Okay let's unpack this first one...the whole visibility game Because frankly..The way buyers find you is just fundamentally fracturing right now.

00:01:11: Oh absolutely!

00:01:12: It's night & day

00:01:13: Right.

00:01:13: I mean Connor Gillivan and Kieran Oliver.

00:01:15: both nailed the core for like a decade.

00:01:19: BB marketers have just played the SEO game.

00:01:22: You know, you optimized to get ranked in a list of ten blue links

00:01:26: right?

00:01:26: The classic Google page one hustle

00:01:28: exactly.

00:01:30: but now we're moving to AEO answer engine optimization.

00:01:34: and the thing with AEO is it doesn't rank you.

00:01:37: It gets you recommended by the AI as this single best answer.

00:01:42: What's fascinating here is the psychological difference for The Buyer.

00:01:45: When you use traditional search, uh...the burden of choice is still completely on you.

00:01:50: You have to click the links do the reading synthesize the info?

00:01:52: Yeah

00:01:53: it's a lot work.

00:01:54: But when A Buyer goes through tool like chat GPT or perplexity the AI does all that synthesis For them and doesn't offer menu.

00:02:00: It offers a verdict.

00:02:01: A

00:02:01: verdict?

00:02:02: That great way To put it.

00:02:03: And your either embedded in that verdict Or just don't exist.

00:02:07: There no page two In an AI Chat response

00:02:11: And I think that's what makes this so incredibly complex for marketers.

00:02:15: Jason Grad actually shared some really wild data from Massive's new API about this.

00:02:20: Oh yeah, the Visibility Checker.

00:02:22: Right they build a visibility check to see what different AI engines recommend and he ran his search for The Best AISDR.

00:02:30: So if a buyer asks ChatGPT... ...the model aggressively pushes tools like Regie & Clay.

00:02:35: Okay it make sense.

00:02:36: But If that exact same buyer types the exact same prompt into perplexity It recommends Zoom info and Agent Frank.

00:02:44: So, I mean how can the best answer be completely different depending on the platform?

00:02:48: Well it

00:02:48: comes down to how these specific models are trained right and what context they actually value, ChatGPT tends to weight its recommendations toward human augmentation narratives.

00:02:58: Oh interesting!

00:02:58: Yeah which aligns perfectly with how tools like Regie position themselves.

00:03:02: but perplexity.

00:03:04: it's built heavily around live data retrieval in citations so its underlying algorithms naturally favor platforms like zoom info that process massive amounts of data.

00:03:13: So basically your AEO strategy can't just be this one-size fits all blanket approach.

00:03:18: Exactly, every single model has its own loyalties.

00:03:21: essentially

00:03:22: Okay I hear the urgency there but i'm gonna play devil's advocate for a second because Amanda Pressner cruiser brought up really compelling counterpoint to all of his panic

00:03:31: Right about the new Google AI guide.

00:03:34: Yeah google published their official guide on ai search and they told marketers to calm down.

00:03:41: They literally claim that optimizing for AI is just SEO.

00:03:46: Like, if your content is good our AI will find it.

00:03:49: Yeah but Amanda pointed out the massive blind spot

00:03:51: there.

00:03:51: It's just SEO.

00:03:52: from Google's perspective they're talking about their own AI overviews which are still built on top of their traditional indexing infrastructure.

00:03:59: But buyers aren't using Google anymore.

00:04:01: That's the whole

00:04:01: point.

00:04:02: Exactly I mean saying a library is same as personal shopper is ridiculous And Amanda shared a study from Aarefs that totally refutes Google's claim.

00:04:11: They looked at the overlap between Google's top ten search results and what AI assistants like ChatGPT actually cite.

00:04:17: Do you remember the number?

00:04:18: It was incredibly

00:04:19: low!

00:04:24: hundreds of thousands of dollars to rank number one on Google, and there is still an eighty-nine percent chance that Chad GPT doesn't even know you exist.

00:04:31: Which is exactly why marketing teams are just scrambling to buy AEO software right now.

00:04:36: but honestly... Even THAT is creating all these new traps!

00:04:39: How so?

00:04:40: Well Salim Malif recently shared his experience testing HubSpot's new AEO tool.

00:04:46: The tool acts like a mirror So it shows how different AI models describe your brand.

00:04:50: But Salim issued this massive warning about people using it.

00:04:54: Oh, because of the prompt suggestion?

00:04:56: Yes.

00:04:56: The tool suggests pumps that you should try to rank for but they sound like old-school rigid keyword foizes.

00:05:03: They don't sound natural conversational questions That a real human would actually type into chatbot

00:05:10: I see.

00:05:10: so if you just blindly optimize for tools suggestions You're optimizing what Celine calls Synthetic demand.

00:05:17: Precisely!

00:05:18: You end up polluting your own site with robotic content that answers questions no actual buyer is ever asking.

00:05:23: So what's the fix?

00:05:24: Salim says you have to bypass a software suggestions entirely.

00:05:28: He pulls actual verbatim questions from human sales, call transcripts and feeds those into his AEO strategy.

00:05:36: Focus on real-buyer language

00:05:38: That focus on human context Is so critical And it goes deeper than just words.

00:05:44: in a web page Garrett Sussman shared this wild experiment that completely blew my mind.

00:05:49: Oh the Gmail

00:05:50: one?

00:05:50: Yes, so he seeded mentions of specific brands into personal gmail accounts meaning these brands were just casually discussed in private personal emails.

00:05:59: right when you then use Google AI mode In that connected account to search for solutions The recommendations for those specific brands shot up by forty six percentage

00:06:08: points.

00:06:08: it is terrifying and brilliant all at once.

00:06:11: It proves AI isn't just crawling public blog posts, it's literally reading personal context!

00:06:16: Yeah...it is using the hidden contexts of a private inbox to shape its public recommendations….

00:06:21: …It completely blurs the line between outbound email and inbound search visibility.

00:06:26: Before we move off visibility though let us save our listeners some time and kill a couple myths

00:06:31: please do.

00:06:32: Jim H confirmed that adding an LLMs.txt file to your website does absolutely nothing, Google confirmed it doesn't work and server logs show no major bots are even looking for it.

00:06:42: Thank you!

00:06:43: And Zach Perix reminded the community about EatEat Experience, expertise, authoritativeness and trustworthiness.

00:06:51: Marketers treat it like a magic ranking factor or code they can hack?

00:06:55: Right just slap a bio on the page you're good!

00:06:57: Exactly.

00:06:58: but its not a magic hack It's quality framework used by human raiders.

00:07:02: You cant trick an AI into thinking your authoritative Just changing your page structure.

00:07:07: So true ok.

00:07:08: so if AEO means buyers are getting single answer recommendations Traditional top of funnel inbound traffic is naturally going to plummet.

00:07:15: Yep volume vanishes

00:07:17: And when marketers see inbound drop, they panic.

00:07:19: They try to make up for that lost pipeline by just forcing demand through massive outbound email campaigns which smoothly brings us to our second theme the reality check hitting AI Outbound and the SDR

00:07:31: function.".

00:07:31: And this is where tension between scale-and-spam really explodes.

00:07:36: SubmitN reported a staggering stat – forty seven percent of AISDR rollouts collapse completely within their first ninety days.

00:07:45: Wait almost half why?

00:07:46: The cause of death is almost always domain reputation.

00:07:49: An AI outreach tool removes all the friction in researching and typing, so a team that used to send fifty careful emails a day

00:08:01: That's a red flag.

00:08:02: Huge one, Microsoft.

00:08:03: three sixty five which hosts like sixty percent of all B to be inboxes monitors these sending patterns.

00:08:09: when their algorithms see an unnatural vertical spike in email volume they instantly flag it as non-human behavior.

00:08:15: It

00:08:15: bounced the emails.

00:08:17: They silently route your entire company's domain straight to the junk folder.

00:08:20: Your dashboard says delivered but no human ever sees it.

00:08:23: Wow

00:08:24: and Eddie Reynolds actually brought the receipts on this exact phenomenon.

00:08:27: You noted that right now per rep outbound volume is up six x across the industry, but reply rates have simultaneously dropped thirty eight percent.

00:08:35: We are literally automating a mess just to get ignored faster

00:08:38: exactly.

00:08:39: we're burning our domains to the ground.

00:08:41: Eddie's point Is that?

00:08:42: Outbound still works But only if you have a rigorous human process.

00:08:46: first.

00:08:47: You need a specific ideal customer profile tightly segmented lists and human inspection.

00:08:52: Yeah Only after that foundation as flawless can you layer AI on top for scale because there's a massive danger in thinking AI can handle the human element of persuasion.

00:09:02: Oh,

00:09:02: absolutely!

00:09:03: Kapildev Aromosi shared a crazy story about a sauce company that tried to go all-in on this — they replaced ten human SDRs with twenty AI agents...

00:09:13: And let me guess… it looked great on paper?

00:09:15: It was a miracle on paper—response times dropped from hours to seconds, outreach volume exploded.

00:09:21: but the moment a deal got emotional or political …the system completely broke.

00:09:25: Because B-to-B buying is inherently emotional.

00:09:28: When a buyer's facing internal budget uncertainty or political risk, the AI can't read the subtext.

00:09:33: it just keeps spitting out feature benefits.

00:09:35: Right

00:09:35: buyers instantly recognize their robotic patterns trust evaporated and deal stalled And Morgan J. Ingram warned about the long term fallout here.

00:09:42: The farm

00:09:43: system right?

00:09:44: Yes Cutting SDRs to save money destroys the farm system for future revenue leaders.

00:09:49: Your Future VP of Sales is currently a twenty-three year old SDR learning how to handle rejection and build empathy on cold calls, you can't promote a bot to VP in ten years!

00:09:59: That's such a crucial point.

00:10:01: so if fully automating this seat as trap…how are best teams actually using that tech?

00:10:07: Prabhupjain and Brendon Short laid out their winning playbook here.

00:10:11: Calling it an AISDR is actually the problem because it implies replacing a seat.

00:10:16: The best teams use AI strictly as an execution layer for the seventy percent of non-selling work.

00:10:22: So, the busywork?

00:10:23: Deep account research CRM updates that kind of thing

00:10:26: Exactly Which frees up humans to spend all their time selling.

00:10:29: Kyle Poyar shared an AI native play that perfectly illustrates this.

00:10:33: He uses Claude not to blast net new leads, but to continuously monitor closed lost deals in the CRM.

00:10:40: Oh

00:10:40: I like that!

00:10:41: Right The AI tracks intense signals Like leadership changes at target accounts.

00:10:45: Then it analyzes old call transcripts To find exact objection killed back then.

00:10:50: And then It drags email.

00:10:52: Yes, it auto-drafts a hyper relevant reengagement Email But and this is key does not hit send Routes it into human rep for review.

00:11:03: That is

00:11:03: brilliant!

00:11:04: Another great use case came from Majovoje, she talked about One Minds right along Superhuman...

00:11:09: Oh the AI sales engineer?

00:11:10: Yeah

00:11:10: it's an AI that acts as a sales engineer on live Zoom calls.

00:11:13: It literally joins The Meeting listens and answers technical Q&A Live in the chat strictly tethered to the company's own documentation.

00:11:21: so doesn't hallucinate..

00:11:23: It totally removes the bottleneck of waiting two weeks For A Human SE To Be Available.

00:11:28: The AE can just focus on strategy.

00:11:30: But you know, executing these advanced plays like AI monitoring intense signals or joining Zoom called it requires a totally different technological foundation.

00:11:38: You can't

00:11:38: just plug that into a ten-year old email tool?

00:11:40: Exactly!

00:11:41: Which brings us to our final theme GTM agents and the system centric architecture.

00:11:47: What does this all mean for the marketing tech stack?

00:11:50: Well, Sandeep Gulati introduced great framework for this He argued.

00:11:54: most teams don't actually have an AI problem.

00:11:56: They've got workflow problems disguised as tools.

00:11:59: Ah, so they buy a thirty dollar license hand it to their team and expect magic.

00:12:04: Pretty much tool-centric teams expect one model like ChatGPT.

00:12:08: do everything.

00:12:09: System centric teams map specific tools to specific workflows based on what they're naturally good

00:12:15: at.

00:12:15: Right but Charlie Saunders issued massive warning here.

00:12:19: He said we have stop using probabilistic AI for deterministic

00:12:23: tasks.

00:12:24: Yes such an important distinction.

00:12:26: Like if a simple Salesforce logic rule can update a campaign status from registered to attended, do not waste AI tokens on it?

00:12:34: The AI might guess right ninety nine times but in the hundredth time it hallucinates at totally weird status.

00:12:41: Right.

00:12:41: so why are many teams getting this wrong?

00:12:43: It comes down into human adoption gap.

00:12:45: Ryan Staley asked this incredibly sharp question Who owns Tuesdays at ten am?

00:12:50: What does that mean like Tuesday morning.

00:12:52: He means who is the specific human dedicated to making sure the sales team actually logs in and uses the new AI workflow at ten a.m on a Tuesday?

00:13:01: The tech works, but enterprises lack a specific role for behavioral change.

00:13:05: Ah

00:13:06: so without that the licenses just gather dust.

00:13:08: well John Millard pointed out That there's also a capability ceiling with legacy marketing automation platforms.

00:13:14: You know MAPs like Older Marketo or HubSpot setups.

00:13:18: he said bolting AI onto a fifteen-year old Lead-centric rules based foundation is just paving the cow path.

00:13:25: It's perfect metaphor, you're taking a winding and efficient dirt trail covering it in concrete.

00:13:30: You can travel terrible paths slightly faster but still a terrible path.

00:13:34: Right legacy platforms were built around single email addresses BDD buying today complex account committees Real time signals

00:13:42: Which is why an AI native architecture has to be what developers call headless.

00:13:47: The back-end data is completely decoupled from the front end interface, so that AI can freely pull from your CRM website and billing software all at once.

00:13:56: So it's not restricted by a rigid software menu?

00:13:59: It could actually coordinate actions across an entire buyer journey!

00:14:02: Exactly… And Michelle Tate really brought this home – she said next advantage won't come using AI to accelerate old workflows...it will come building entirely new agentic operating models.

00:14:12: We have to completely rebuild the workflow

00:14:15: We do And that leads perfectly into a final thought for you to mull over, drawing on some great insights from Cone Stam and Dornrata.

00:14:23: AI isn't just a new tool to bolt onto the marketing department.

00:14:27: it is The New Operating System of Growth.

00:14:29: Yeah That's A Powerful Way To Look At It!

00:14:31: The real competitive edge belongs to the CMOs who realize that THE NEW AI PLAYBOOK ISN'T GOING TO BE WRITTEN BY THE SOFTWARE VENDORS.

00:14:39: IT'S GONNA Be Written By Teams Right There In Trenches Who Are Brave Enough To Rethink Their Entire Customer Journey

00:14:45: Absolutely.

00:14:46: Well said!

00:14:47: If you enjoyed this episode, new episodes drop every two weeks.

00:14:50: Also check out our other editions on account-based marketing field marketing channel marketing martech go to market and social selling.

00:14:57: Thank you so much for joining us on this deep dive And don't forget to subscribe.

00:15:00: see ya next time.