Best of LinkedIn: AI in B2B Marketing CW 09/ 10

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

00:00:07: In calendar weeks nine and ten.

00:00:09: Frennis is a B-to-B market research company helping enterprise marketing teams sharpen their strategies and outreach with customer segmentation ideal customer profiles and deep dives Customer needs analyses And buying center insights

00:00:23: And welcome to the deep drive.

00:00:25: We are super excited to have you with us today because honestly, the landscape right now is wild.

00:00:31: Oh absolutely like.

00:00:32: what if I told you that?

00:00:33: The average bdb marketing team is sitting on software stacks where You know fifty three percent of the licenses.

00:00:39: just sit there completely unused.

00:00:40: That is i mean that's Just burning money Right

00:00:43: exactly and they're still missing their revenue quotas.

00:00:45: we've spent this entire last decade buying tools To make selling easier And in a lot of cases, we've basically just built a really expensive high-speed spam machine.

00:00:54: Yeah which is why today for this deep dive We're looking at the top AI trends in B to be marketing that are actually surfacing on LinkedIn right now and Just to set expectations For you listening?

00:01:06: We Are strictly moving past The whole hype phase Today.

00:01:10: Right no crystal ball predictions here exactly

00:01:12: know?

00:01:12: You Know what I might do In five years stuff were Looking At the operational reality On the ground today The real workflows enterprise teams use to drive revenue.

00:01:22: And

00:01:22: the hard truths too, like where these automated systems are frankly just completely breaking down?

00:01:28: Totally.

00:01:29: which actually brings up this really interesting shift.

00:01:31: and how we even view the tech.

00:01:33: because you know previously when a new marketing tool dropped it was like buying a power drill for your workshop.

00:01:40: You plug it in learn the settings and build that exact same birdhouse faster right

00:01:45: A localized efficiency game, but looking at the content from weeks nine and ten We aren't just buying new drills anymore.

00:01:51: The entire factory floor is like being redesigned From the ground up.

00:01:55: it's a total paradigm shift.

00:01:56: It really is its moved from being this fun novelty-like Drafting a quick social post to be literal load bearing infrastructure.

00:02:06: Yeah.

00:02:06: And Ada Varensman this really compelling point about that shift.

00:02:09: on linkedin.

00:02:10: he argues That what?

00:02:11: He calls Vibe marketing is becoming its own distinct lane now.

00:02:16: Vibe

00:02:16: Marketing, I love that term!

00:02:17: Right it's great but he says if you're still doing marketing the twenty-twenty one way You're gonna feel the friction incredibly fast because in the past The bottleneck was always production

00:02:28: right.

00:02:29: like how fast can the designer make this graphic?

00:02:31: Exactly or editing the video.

00:02:33: But today a dev notes that production is basically instant.

00:02:37: So the new bottlenecks are taste, velocity and distribution.

00:02:40: That's really interesting.

00:02:41: so it is more about how quickly you can iterate on a creative concept until actually clicks with audience

00:02:47: Precisely.

00:02:48: And he points out that if your tech stack relies on traditional tools even popular ones like Canva or Capcut but without deep native AI integration You're essentially doing manual labor While

00:02:58: everyone else has an automated factory.

00:03:00: Exactly Like he's running an AI native stack with tools like NanoBanana for images, Higgs Field for video, whisper flow of voice to text.

00:03:09: The point isn't just making art.

00:03:11: It is typing a prompt getting twenty variations in ten seconds picking the one that matches your brand vibe and shipping it before traditional team even opens a blank file.

00:03:22: Yeah, that kind of velocity completely changes the financial math of a campaign.

00:03:26: And actually Dimitrov shared a real-world example of this.

00:03:30: that just perfectly illustrates the impact.

00:03:32: Oh!

00:03:32: The e-commerce one?

00:03:33: So he watched this e-commerce store owner burn through like eight thousand dollars on blanket ads and they had zero clue where the actual sales were coming from

00:03:44: just spraying budget everywhere.

00:03:45: Pure guesswork, so they plug an AI driven marketing automation stack into their ad accounts and not to plan a strategy for next quarter but to apt the very next day?

00:03:56: I saw that post in.

00:03:57: what really stood out to me wasn't that the AI just magically wrote a better headline.

00:04:01: Right it was about creative.

00:04:03: No It's the mechanism.

00:04:04: The AI is just ruthless with the budget because you know A human media buyer might check the campaign twice a week

00:04:12: Maybe if there on top of Right.

00:04:14: The AI monitors the return on ad spend minute by minute, it literally cut the dead campaigns before breakfast.

00:04:21: That is wild.

00:04:22: And then dynamically redirects that spend to three percent of ads doing eighty percent.

00:04:27: Yeah and Dimitar mentioned its stopped pouring money into what he calls Hopium audiences.

00:04:32: Hopium Audiences, that is such a good phrase!

00:04:35: It really

00:04:35: is.

00:04:36: it's those demographics at a human marketer desperately hopes will buy because they fit some theoretical persona.

00:04:42: but the live data shows They just aren't clicking

00:04:45: right.

00:04:45: and the result ad spend dropped by forty two percent in week one Forty-two

00:04:49: percent And revenues stayed completely flat

00:04:52: which means profit margins absolutely skyrocketed.

00:04:55: Yeah as AI doesn't have an emotional attachment to a campaign its Just ruthless with waste.

00:05:00: Yeah, humans really struggle to kill their darlings like that.

00:05:03: But you know if AI is completely upending how we execute ad spend it's also totally changing How our brands are discovered organically.

00:05:11: Oh for

00:05:11: sure organic search Is completely different now.

00:05:14: Right

00:05:14: and Carlo Alberto Kiman shared this framework That dissects This.

00:05:18: He argues most companies Are failing at AI Search visibility Because they ignore the sequence Of how machines actually learn about a brand.

00:05:25: You

00:05:25: have three stages.

00:05:26: he breaks It down so well.

00:05:28: First you have traditional SEO search engine optimization.

00:05:31: Right, the classic stuff?

00:05:32: Exactly because that enables initial discovery The machine has to know you exist in first place.

00:05:37: Then second stage is AEO answer Engine Optimization

00:05:41: Which where structure site data?

00:05:43: so AI actually comprehends context not just keywords

00:05:47: Spot on.

00:05:48: And finally third stage is GEO generative engine optimization.

00:05:53: Okay, let's actually pause here and clarify GEO because I feel like that term is getting thrown around a lot right now.

00:05:58: How do you actually optimize for a generative Engine?

00:06:02: So Geo is really all about enabling trust.

00:06:05: When a user asks ChatGPT for say, a vendor recommendation the AI synthesizes info to give one single authoritative answer.

00:06:13: Right it doesn't just give you list of links.

00:06:15: Exactly

00:06:15: so.

00:06:16: be that one answer.

00:06:16: You need third-party validation strong brand mentions across the web and really dap authoritative content.

00:06:23: Carlos' whole point was that if you break this sequence, like just try to skip straight to GEO without doing the SEO and AEO work-nothing compounds.

00:06:31: You can't skip foundational steps!

00:06:33: And SWATI Pollywall added a super crucial layer on LinkedIn too.

00:06:38: Oh yeah...the sea of AI visibility.

00:06:40: Right..she says the most important C isn't content anymore.

00:06:43: it's consistency

00:06:44: which makes total sense.

00:06:45: Yeah,

00:06:46: meaning consistency across how you frame industry problems your specific language repeating the same case studies because an AI system is constantly cross-referencing data across the entire web.

00:06:56: it doesn't reward occasional brilliance

00:06:58: exactly!

00:06:59: It rewards repeated clarity.

00:07:01: if your website says one thing and your LinkedIn says another in your PRs doing something else the AI gets confused and just recommends your competitor who actually has a unified message.

00:07:12: And achieving that absolute clarity requires some really tactical adjustments.

00:07:16: Marcus shared, he shared a super practical pit for them.

00:07:19: Oh I love this one!

00:07:20: Yeah He says you need to put clear bulleted TLDR like too long didn't read.

00:07:26: summary at very top of every single blog post.

00:07:30: Yes

00:07:30: And he suggests replacing those traditional generic H-II header tags with very specific questions.

00:07:36: So like instead of a header that just says pricing models, you change it to.

00:07:40: how much does Enterprise B-to-B software cost in twenty twenty six?

00:07:44: Exactly because why?

00:07:45: It still serves legacy Google SEO but it absolutely spoon feeds large language models exactly what they crave.

00:07:52: right Because

00:07:52: LLMs are literally built on question and answer.

00:07:55: training data

00:07:55: Precisely You serve the algorithm The exact format it needs To retrieve your info.

00:08:00: Structure really is everything here, and the stakes for getting this right are massive.

00:08:05: Steve Armenti actually shared some data from CallRail that blew my mind.

00:08:09: The inbound calls data.

00:08:11: Yeah

00:08:11: They analyzed twenty million inbound calls and found that AI search tools like ChatGPT, Perplexity, Claude Ball are driving actual phone calls to real businesses.

00:08:22: We're not just talking about clicks on a landing page anymore.

00:08:24: Exactly!

00:08:25: ChatGPP is driving ninety percent of those AI referred calls especially healthcare and automotive because buyers want an immediate answer.

00:08:35: I think the metaphor here is so important, like for two decades Google was basically a library.

00:08:40: You walk in ask a question and it hands you ten aisles of books to wander through.

00:08:44: You're just on research mode opening twenty tabs?

00:08:46: Yeah exactly but an AI search engine as a concierge...you asked a question..and it hands your exact tool to solve that problem.

00:08:53: right then there The buyer intent....is completely different.

00:08:57: It's so much closer to the actual point of purchase.

00:08:59: Totally, if you aren't a tool that concierge hands over You functionally don't exist in that cycle.

00:09:05: Yeah but okay I have to challenge this structure A little bit.

00:09:08: Go

00:09:08: for it

00:09:08: Because If your listening To This Right And you decide to follow Marcus Sheridan's advice right?

00:09:13: He put TLDRs everywhere.

00:09:15: Turn all your H-twos into questions.

00:09:17: Optimize perfectly For The Machine.

00:09:19: If every single B-to-B brand does this, aren't we risking just total homogenization?

00:09:24: Well for sure.

00:09:24: Like if everyone sounds exactly the same to the algorithm using the exact same format how does a brand actually stand out?

00:09:32: That is the critical risk and I think the answer really comes down to substance over syntax.

00:09:38: You fight that homogenisation with proprietary data.

00:09:42: Okay meaning what?

00:09:43: Meaning... But your answer relies on generic tropes and their answer includes original research or a highly specific case study.

00:09:53: The LLM will favor the depth in originality.

00:09:56: Ah,

00:09:56: okay.

00:09:57: So you stand out through unique substance not just this shell correct?

00:10:00: The format Just ensures that machine can actually read the value You're providing.

00:10:04: That makes a lot of sense And you know if marketing's job is to use these tools To generate that intent it begs the question how was sales Actually acting on

00:10:13: right?

00:10:13: If the concierge hands them your name What happens next?

00:10:16: Exactly,

00:10:17: which

00:10:17: brings us to the second big theme we saw.

00:10:20: The reality of AI-SDR sales development reps and automated workflows... ...the promise here is just incredibly alluring!

00:10:28: Oh it's the Holy Grail.

00:10:30: Roman Charney and Vanessa Ponce shared a concept that completely reframes outbound.

00:10:35: They argue the core problem with outbound isn't usually the message itself, it's

00:10:40: timing.

00:10:41: Because traditional prospecting is human scraping a list of five hundred contacts blasting emails or just praying someone was in buying window but B to B buyers only by when the status quo changes.

00:10:52: Like a new executive hire, or funding round?

00:10:54: Exactly!

00:10:55: So instead of relying on human timing they've built an autonomous system using a platform called Gojiberry AI.

00:11:02: Okay so how does that mechanism work?

00:11:03: Rather than blasting static lists The system constantly monitors data streams like linked in API for specific trigger events.

00:11:10: Yeah...so at this moment a target account announces say A Series B Funding Round the AI agent moves It cross-references the new hire against the ICP, analyzes recent activity to draft a highly contextualized message and launches the outreach instantly.

00:11:26: Autonomously.

00:11:26: Striking

00:11:27: while the iron is hot.

00:11:28: Dylan Power shared very similar operational win actually using Claude Coworker.

00:11:33: Oh I saw that!

00:11:33: The data cleanup one.

00:11:34: Yeah

00:11:35: he treats Claude like research co-pilot for outbound because we all know how messy CRM data gets.

00:11:41: it's always disaster.

00:11:43: So instead of an SDR spending four hours doing VeloCups on a spreadsheet Dylan just feeds the raw CSV into Claude.

00:11:50: He prompts it with their exact ICP criteria and Claude autonomously cross-references.

00:11:54: The rose surfing is the real decision makers and pulls contextual signals.

00:11:58: That's

00:11:58: amazing.

00:11:59: right by removing that manual drag his team closed.

00:12:02: six hundred twenty thousand dollars in Just forty five days.

00:12:05: Wow

00:12:06: because the operators were freed up to actually do the human judgment part

00:12:08: exactly Closing the conversations, but we do need to inject a serious reality check here.

00:12:13: Please do

00:12:14: because AISDRs are not just a magic button.

00:12:17: you flip on.

00:12:18: Jason M Lemkin and Amelia LaRoute from Sauster shared their experience running over twenty of these AI agents

00:12:24: right?

00:12:24: And they booked a ton of pipeline didn't they?

00:12:26: They did Over two point seven five million dollars Yeah.

00:12:30: But Jason was incredibly transparent about the hidden operational costs.

00:12:34: Time maintenance

00:12:35: The massive maintenance.

00:12:36: Every single agent requires roughly thirty days of hands-on training before it could be trusted.

00:12:42: Thirty

00:12:42: days per Agent.

00:12:44: Yep, you have to manually review first thousand outputs because they hallucinate.

00:12:49: They make these logical leaps a human never would.

00:12:52: Managing those agents currently consumes thirty percent their chief AI officer's time.

00:12:56: That is seriously heavy.

00:12:58: lift.

00:12:58: just get them functional

00:12:59: It is.

00:13:00: And Scott Martini's pointed out an even deeper architectural flaw in how we deploy these agents right now, he says AISDRs are stuck in the SAAS paradigm.

00:13:10: Wait what does it mean by that?

00:13:11: He means We treat them like software filters.

00:13:13: You know you configure the parameters run the LLM and track open rates But they fail to do what a skilled human SDR does naturally, which is provide strategic feedback.

00:13:24: Ah okay

00:13:25: Like if the HumanSDR makes fifty cold calls and gets hung up on forty times.

00:13:29: They notice patterns!

00:13:30: They go into marketing saying Hey this positioning lands with VPs but directors are totally confused.

00:13:35: Right Humans hypothesize

00:13:37: Exactly.

00:13:38: AI agents right now just blindly execute the configuration.

00:13:42: They don't form hypotheses about why campaign is failing.

00:13:44: I am so glad you brought that blind execution problem because we have to be so careful not just industrialize our bad habits.

00:13:52: Kevin N posted a great observation about this, he noted that the loudest voices pushing AISDRs are the ones claiming we should stop hiring humans.

00:14:01: but you know outbound usually fails because of bad inputs.

00:14:04: Bad data weak offers

00:14:06: Exactly zero real buying signals.

00:14:09: If your underlying strategy is just a generic spray-and-pray approach and you bolt an AI agent onto it, You aren't innovating.

00:14:16: You're just automating failure at light speed Just

00:14:18: creating the biggest spam problem ever.

00:14:20: Exactly Which is why The underlying data layer... ...the infrastructure matters way more than whatever flashy AI wrapper you buy.

00:14:28: And frankly for most BtoB teams right now.. ..The tech stack Is A COMPLETE disaster.

00:14:33: Oh It's a sprawling mess!

00:14:35: Which brings us to our third theme tools and the tech stack.

00:14:40: Orin Greenberg shared some truly jarring stats on this architectural debt.

00:14:44: Let's hear it

00:14:45: The average B to B sales rep right now is forced use eight different tool stitched together.

00:14:50: just do their job.

00:14:51: Eight

00:14:51: Tools?

00:14:52: That's ridiculous.

00:14:53: And even worse, fifty-three percent of enterprise software licenses are sitting totally unused.

00:14:59: We spent a decade buying specialized tools, and we've just created a fragmented nightmare where data is trapped in silos.

00:15:06: And were seeing the market aggressively react to that right?

00:15:08: Yeah.

00:15:08: Olivier Medell noted looking at twenty-twenty six The question in enterprise boardrooms isn't which AI tool should be buy next.

00:15:15: No it isn't.

00:15:15: It's Which eighty percent of our A.I.

00:15:18: Tools are going kill this year.

00:15:19: What He seen?

00:15:20: enterprises actively consolidating from like fifteen to twenty five disparate vendors down three to five core platforms, fragmented stacks are literally a board level risk now.

00:15:31: Well yeah if you have disconnected agents making conflicting decisions based on silo data... You lose control of your go-to market motion!

00:15:39: Exactly so.

00:15:40: If hoarding tools is the problem what does smart stacking actually look like?

00:15:45: Spencer Perrick shared his story.

00:15:46: that illustrates this perfectly.

00:15:48: A founder emailed him in total frustration.

00:15:51: The Founder said we HubSpot, Clay for Enrichment Smart Lead Apollo and AISDR.

00:15:58: And we are still completely missing quota.

00:16:00: Let me guess it was not a software problem

00:16:03: Not at all.

00:16:04: It took Spencer twenty minutes to audit the system and find the root cause.

00:16:07: The founder's ideal customer profile –the criteria of feeding these expensive tools–was just a broad TAM filter….

00:16:14: …a total addressable market-filter.

00:16:16: Right Their target was simply BtoB SaaS companies doing fifty in five hundred employees.

00:16:21: That's not an ICP, that is just a demographic reality.

00:16:24: Exactly!

00:16:25: It was way too broad.

00:16:26: So Spencer audited their historical closed one deals and found the hidden signal.

00:16:31: Their best clients were specifically series A sauce companies that had recently hired their very first head of sales.

00:16:37: Ah because founder led sales motion hit a ceiling

00:16:40: Precisely so they stopped blasting the whole TAM.

00:16:43: They rebuilt system to trigger outreach within fourteen days of ahead-of-sales job posting And reply rates jumped from one point eight percent to fourteen point three percent.

00:16:53: That's

00:16:54: incredible, so the tools were fine?

00:16:55: The strategic signal was just broken.

00:16:57: right.

00:16:57: AI should amplify human strategy not replace it and Nadia Cumentis shared a framework for this balance.

00:17:04: She calls at the ten eighty-ten rule.

00:17:06: Oh

00:17:06: I like this one.

00:17:07: It's So practical!

00:17:08: The first Ten percent of any workflow has To be Human Strategy setting the brief defining pain points crafting the prompt.

00:17:15: the middle Eighty percent is the AI doing the heavy lifting

00:17:17: processing data drafting assets yeah.

00:17:20: And the final ten percent is human review.

00:17:22: Critical evaluation, editing for tone,

00:17:25: fact-checking.".

00:17:25: So if you're launching a webinar... You don't just tell the AI right an invite!

00:17:29: That first ten percent IS YOU inputting specific industry shifts and proprietary data.

00:17:34: Exactly.

00:17:35: Skip that first ten per cent?

00:17:37: You get generic spam?

00:17:38: Skip the final Ten Percent?

00:17:39: You publish hallucinations AND damage the

00:17:41: brand?!

00:17:42: The HUMAN has to be in the driver's seat.

00:17:44: I love that framework but even when an organization gets THAT rule RIGHT There's this massive operational risk hiding in plain sight.

00:17:53: Nicole Leffer shared an observation about team dynamics that every marketing leader needs to hear...

00:17:58: Oh, the pipeline issue!

00:17:59: Yeah she calls it The In-House AI Lead To Independent AI Consultant Pipeline.

00:18:04: This is

00:18:05: such a dangerous pattern right now.

00:18:06: I pray.

00:18:07: Here's how It Happens.

00:18:09: A company has one naturally curious marketer.

00:18:12: They let this person build all their custom workflows, the prompt libraries and API connections.

00:18:17: they become the AI whisper for whole department.

00:18:20: I've seen this play out

00:18:21: And inevitably that marketer realizes there market value has skyrocketed.

00:18:27: So they leave to become a highly paid consultant.

00:18:30: And the company is suddenly left with a bespoke infrastructure that absolutely no one else understands,

00:18:36: and then whole system just collapses

00:18:38: exactly which raises huge organizational question for you listening.

00:18:43: are You're buying isolated tools so you look innovative or Are you actively building resilient marketing operating systems?

00:18:51: A true system means workflows are documented multiple people understand the APIs and it serves a unified strategy.

00:18:58: You can't have a single point of

00:18:59: failure.".

00:18:59: Exactly, so we've covered this shift to infrastructure autonomous workflows cleaning up the tech stack.

00:19:05: but as we look ahead I want leave you with final thought building on an insight from Ryan Law.

00:19:09: Yeah, this is a really profound point.

00:19:11: Because we've spent this deep dive talking about automated factories for content.

00:19:15: but Ryan points out that very soon tech companies will fully automate the mechanics of content creation.

00:19:21: it Will be seamless

00:19:22: indistinguishable from human writing

00:19:24: right.

00:19:25: and when that happens When an automated content engine?

00:19:27: Is just unremarkable.

00:19:28: table stakes.

00:19:29: What is your competitive advantage?

00:19:32: The most valuable asset you have won't be your tech stack.

00:19:35: Everyone will have the exact same stock, sure!

00:19:37: Your ultimate moat will be your personal brand... ...your spiky highly subjective opinions and authentic human trust built with audience.

00:19:45: You can automate an email but cannot automate trust.

00:19:48: If you enjoyed this episode new episodes drop every two weeks.

00:19:52: Also check out our other editions on account-based marketing field marketing channel marketing marktech go to market & social selling.

00:20:00: Thank you so much for joining us on this deep dive.

00:20:02: Don't forget to hit subscribe, So don't miss our next exploration.

00:20:05: and as You review your systems this week remember to look at the blueprint not just The drill.

00:20:10: we'll see you next time.