Best of LinkedIn: AI in B2B Marketing CW 31/ 32

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

00:00:07: In calendar weeks thirty one and thirty two.

00:00:09: Frennis is a b to be market research company that supports enterprise marketing teams in unlocking the full potential of their customer data with the help of A I. so um what have i told you?

00:00:21: That By The Time we hit like twenty twenty eight your absolute biggest Most lucrative client might not even Have a pulse?

00:00:29: super jarring thought, right?

00:00:31: But honestly that is exactly where the data's pointing us today.

00:00:34: It

00:00:34: totally changes everything.

00:00:36: so The mission for this deep dive is to really unpack this massive fundamental shift that's happening right under our feet.

00:00:43: Yeah, the center of gravity in B-to-B marketing is radically moving Right.

00:00:46: we're going to explore how brands basically have to stop optimizing To be just you know found and start optimizing to be cited by machines

00:00:54: Exactly And will look at how top tier sales teams are like literally rebuilding their entire infrastructure around encoded AI judgment.

00:01:02: yeah NY marketing leadership Is frankly kind of scrambling right now.

00:01:05: how to drive this thing before they crash it.

00:01:07: The stakes have really never been higher, I mean we are transitioning out of a reality where you controlled your digital real estate right

00:01:13: the good old days yeah

00:01:15: You used to optimize your website and then you ranked on Google.

00:01:17: It was very direct relationship but now We're entering This world were completely dependent On How an autonomous black box system, interprets your brand.

00:01:28: It fundamentally rewrites what top of funnel marketing even is.

00:01:32: because let's be real the old rules of search are just dead

00:01:36: totally dead.

00:01:37: it's no longer about whether you can keyword stuff Your way to page one.

00:01:41: today The only question that actually matters Is does this AI model trust?

00:01:45: You're brand enough To cite you in its response to a buyer

00:01:50: And how it builds that trust is where things get incredibly complex.

00:01:54: Kyle Atwater Morley recently unpacked the mechanism behind this and its a really critical detail for anyone listening.

00:01:59: Okay, break down for us!

00:02:00: So AI platforms are waiting what the wider web says about your business far more heavily than what you're business has itself.

00:02:07: Oh wow okay Yeah

00:02:08: You can have the slickest homepage in world Claiming Your The Leading Solution In Your Space.

00:02:13: But the AIs out there looking For External Validation.

00:02:17: Kyle introduces this concept of co-occurrence as the new currency next to a specific concept across the internet.

00:02:41: Okay, so if you sell say enterprise cloud security The AI model is scraping third-party sites industry directories reddit forums independent tech blogs

00:02:50: news articles all that stuff right

00:02:52: All of it.

00:02:53: and if your company's name consistently appears in the exact same paragraph as the phrase Enterprise Cloud Security out there in the wild?

00:03:03: Yeah,

00:03:03: every single time that happens it strengthens the machine's mathematical confidence.

00:03:07: That you are in fact The answer to that specific problem.

00:03:11: I mean i get the logic but doesn't that just invite a whole new era of spam?

00:03:16: Like if i'm a marketer can i Just go buy a thousand cheap press releases To force my brand name to sit next those keywords?

00:03:24: You could try But the models Are actually actively guarding against which brings us to a massive risk that Stephanie Amini highlighted.

00:03:32: The AI is highly, highly sensitive to inconsistency.

00:03:37: What do you mean by inconsistence?

00:03:39: Well if your website tells one story And then your LinkedIn company page highlights a slightly different value proposition.

00:03:46: Your crunch-based profile hasn't been updated since your series A three years ago, right?

00:03:50: Which happens all the time

00:03:52: exactly?

00:03:52: and maybe you're CEO uses entirely Different jargon on a podcast.

00:03:56: when all that happens The AI detects of fractured identity.

00:04:00: Oh

00:04:00: That makes sense.

00:04:01: It's basically running a background check on you and if your story changes depending On which referenced the investigator calls You just fail to check

00:04:07: you don't get the job.

00:04:08: that is a brilliant way to frame it.

00:04:10: And in the world of large language models, our fractured identity equals a catastrophic drop in

00:04:16: confidence.".

00:04:17: Yeah?

00:04:17: If the model isn't highly confident and exactly what you do... It doesn't guess!

00:04:22: ...it just omits you.

00:04:23: You vanish from output entirely

00:04:25: Which is terrifying for brand.

00:04:28: But here's my issue with this whole paradigm.

00:04:30: if we're relying on these black box models to validate us How do we even measure if our strategy is working?

00:04:37: That's the big question.

00:04:39: Right,

00:04:39: like how do you track a metric that just changes its mind?

00:04:42: You've hit on an exact crisis happening in SEO departments right now.

00:04:47: Peter Rota issued pretty stark warning about this.

00:04:50: He went and tested current AI visibility ranking tools And found they are deeply flawed.

00:04:57: Flawed how?

00:04:58: Like the data is wrong!

00:04:59: Well, because AI is non-deterministic meaning it can generate a totally different answer to the exact same question.

00:05:05: He found that a brand's position in an AI response Can swing by as much as ten spots on the exact Same prompt.

00:05:11: wait really just from asking The same thing twice.

00:05:14: yes Just asked a few minutes apart.

00:05:16: That's

00:05:16: wild if the numbers change every time you hit refresh.

00:05:19: That's not a performance metric.

00:05:20: That's just a random number generator

00:05:22: Precisely Which is why Zohi Mostafa is arguing that we basically need to torch the traditional SEO dashboard entirely.

00:05:29: Torch

00:05:30: it all!

00:05:30: I love it, so what's the alternative?

00:05:32: He suggests this new five-step pipeline to measure AI visibility and It requires a totally different mindset.

00:05:40: Step one can the AI even find you

00:05:42: right foundation level?

00:05:43: step two Can it retrieve you when a relevant query actually happens?

00:05:47: okay fine in retrieve.

00:05:48: step three does it actually cite you as evidence?

00:05:52: Step four, and this is the really big one.

00:05:54: What's your share of authority?

00:05:56: Share of authority, meaning what?

00:05:57: Exactly.

00:05:58: Meaning when it talks about the topic are you the central focus or just a tiny footnote buried beneath your biggest competitor?

00:06:06: and finally step five does the AI actually recommend as this solution to buyer

00:06:11: Man that is much heavier lift than asking did we get click?

00:06:15: Oh

00:06:15: absolutely

00:06:16: What strikes me?

00:06:17: those steps?

00:06:18: how fragmented these models really is.

00:06:21: We tend treat AI like its one single giant brain But it's really not.

00:06:26: Not at all.

00:06:27: Julian Scho shared this fascinating experiment that just perfectly proves this, a research team spun up two websites completely filled with AI generated slop.

00:06:37: Just thousands of nonsensical low quality articles All published in a single day.

00:06:43: A classic spam tactic.

00:06:44: So how

00:06:45: did the search engines handle?

00:06:46: Well Google's traditional search engine worked exactly as it should.

00:06:50: It identified the slop and completely de-indexed sites.

00:06:54: They vanished from traditional search pages.

00:06:56: Right, making the system work.

00:06:57: But

00:06:57: here's the crazy part.

00:06:58: The AI answer engines reacted completely differently Even though Google kicked them out.

00:07:03: Citations for this AI slump in AI Answer Engines actually rose seventeenfold.

00:07:09: Wait!

00:07:09: Seventeenfold Just based on the sheer volume of content.

00:07:13: Yeah, just volume and most those citations came from Microsoft Copilot.

00:07:17: it completely proves that human visibility in traditional search an AI visibility.

00:07:21: in these new engines they're playing by two totally different sets of rules

00:07:25: And even within the AI engines The rules change depending on whose model you are actually looking at.

00:07:30: Oh for

00:07:30: sure

00:07:30: Andrew Warden shared some SEMrush data.

00:07:33: That puts a hard number On this fragmentation.

00:07:36: We know Google AI overviews cite about sixty-four percent of the brands they mention.

00:07:40: Okay,

00:07:40: sixty four percent.

00:07:41: But Gemini which remember is built by The Exact Same Company cites only thirty percent Of the Brand's it mentions.

00:07:48: Wow

00:07:49: Yeah You aren't just playing a different game than traditional SEO you're Playing A totally Different Game on Every Single Platform.

00:07:56: It's honestly exhausting just thinking about it.

00:07:59: Actually, you know if you want to make sure your staying ahead of these rapid week by weeks shifts and how AI is reshaping marketing.

00:08:06: Make sure you hit the subscribe button right now

00:08:08: Seriously?

00:08:09: You don't wanna miss it.

00:08:10: Yeah!

00:08:10: Its easiest way to ensure that you dont' miss our future deep dives because as we're seeing this playbook being rewritten literally AS WE SPEAK

00:08:17: IT REALLY REALLY IS.

00:08:19: Now lets follow the logic of Zohi Mestofas pipeline for a second.

00:08:23: Let's say you nail those five steps.

00:08:25: You've secured visibility, the AI trusts you and it finally recommends you.

00:08:30: We pop the champagne right?

00:08:32: Not quite because a recommendation from a machine doesn't automatically put money in the bank.

00:08:37: Fair point How do you convert that AI citation into actual closed one revenue?

00:08:43: Yeah This moves us directly into the execution layer.

00:08:46: go to market in sales.

00:08:47: Okay let's get into.

00:08:48: And the shift here is profound.

00:08:50: We are moving away from using AI as a basic automation tool, where it just like sends emails faster to using it as an encoded layer of sales judgment

00:09:00: and encoded Layer of Sales Judgment?

00:09:02: I mean, I love the sound of that but have to say am a bit skeptical when i look at the market.

00:09:06: most A.I.

00:09:07: sales development reps The A.i SDR rollouts.

00:09:11: they're flaming out there annoying buyers and failing to book meetings.

00:09:15: Victor Adifie actually addressed this exact phenomenon.

00:09:19: He argues that it is almost never a failure of the software itself.

00:09:23: It's a catastrophic underinvestment by the leadership team.

00:09:25: How

00:09:25: so?

00:09:26: Well

00:09:26: companies are buying an AI tool and treating a capability up.

00:09:29: they should replace A hundred thousand dollar human hire like its a twenty nine-dollar month plug in play subscription.

00:09:34: Oh wow, yeah They turn on give it a super generic prompt and just expect magic.

00:09:39: So how do you actually encode judgment then?

00:09:41: Because, obviously can't tell an AI go be a great salesperson.

00:09:44: Exactly!

00:09:45: Victor says the secret of winning teams is deep-deep architectural work.

00:09:50: They analyze hundreds hours call recordings.

00:09:53: they identify one specific Hyper nuanced thing.

00:09:57: their absolute best human rep does better than anyone else like what.

00:10:01: maybe it's how they connect a really specific pain point to A very niche feature of your product.

00:10:07: then They teach the AI to replicate that exact logical pathway at scale.

00:10:34: Yeah,

00:10:35: and this wasn't just like a Chrome extension they downloaded.

00:10:38: It's a foundational piece of their workflow now.

00:10:40: And the numbers coming out of that are staggering.

00:10:44: Account executives who use Kai produce twice the sales activity

00:10:48: which is great But activities just noise if it doesn't actually convert

00:10:51: exactly.

00:10:52: but The real metric is that?

00:10:53: They create twenty six percent more revenue opportunities and close thirty nine percent More deals compared to the weeks.

00:10:59: They don't use the platform.

00:11:00: A thirty-nine percent bump in closed deals.

00:11:02: I mean, that isn't just a marginal efficiency game That's a completely different tax bracket for those reps

00:11:07: totally.

00:11:07: But you know before everyone listening pauses this deep dive to go buy an AISDR.

00:11:13: We do need to look at the ceiling of this technology because automation has hard limits Especially if your only competitive advantage is that?

00:11:20: You just have access to data right.

00:11:22: The ramp story.

00:11:23: yes

00:11:24: Elric Leg-Laura showed a really revealing case study about this.

00:11:27: Ramp, the finance automation company actually shut down its highly successful AISDR system back in December.

00:11:35: Yeah they had this internal system called OATs The Outbound Automated Team and for years it was an absolute juggernaut.

00:11:43: OATS drove thirty percent of their entire sales pipeline.

00:11:47: So if it's driving almost a third your pipeline why on earth would you turn off?

00:11:51: Because the automated phone just plateaued.

00:11:54: They realized that by twenty-twenty five pretty much every single one of their competitors had access to the exact same data sources and the exact Same Automated outbound products,

00:12:04: so The differentiation completely vanished.

00:12:07: exactly if your mode is Just we send personalized AI emails faster than the other guy you don't actually have a moat.

00:12:14: Ramp realized they sell complex financial products that require deep trust and real human conversations to close.

00:12:21: Right So, They shut down the bot layer but... And this is the genius part!

00:12:25: ...they didn't throw out infrastructure.

00:12:27: They kept underlying snowflake data warehouse built for it.

00:12:31: Ah

00:12:31: so they shifted technology from replacing SDRs to arming them instead?

00:12:36: Exactly, instead of an autonomous bot just spamming in inbox that massive snowflake data infrastructure now sits directly on the human rep screen.

00:12:44: That's

00:12:45: smart!

00:12:45: Yeah it gives the human instant real-time context like highlighting that a prospect just installed three competing tools so the human can have a vastly smarter highly trusted conversation...

00:12:57: ...that is a brilliant strategic pivot and it perfectly illustrates a core argument from Max Foster.

00:13:03: he noted.

00:13:04: then AI tool say Claude only becomes truly useful for a sales team the minute you stop treating it like a conversational chatbot.

00:13:11: Right, You have to stop talking about it as if its smart interns sitting at a typewriter?

00:13:15: Yes!

00:13:15: Exactly!

00:13:17: MaxSense needs being integrated layer... ...you need specific plug-ins for account research.. ..you need direct connectors feeding into your existing tech stack Like HubSpot or ZoomInfo

00:13:26: Making It Seamless.

00:13:28: The AI needs to run on top of your data autonomously pulling context in the background rather than sitting In a separate browser tab waiting for a human to type out a massive prompt every single time they want To send an email.

00:13:39: it just has to be baked into the workflow.

00:13:42: But as we're marveling at all this efficiency.

00:13:44: We have to ask the uncomfortable question if Your core go-to market strategy is fundamentally flawed isn't ai?

00:13:51: Just going to help you fail At the speed of light?

00:13:53: that Is the million dollar questions.

00:13:55: yeah our teams buying Ai to fix a fundamentally broken pipeline, or are they just automating the pursuit of dead deals?

00:14:03: Because generating a thousand highly personalized perfectly encoded emails for a product that nobody actually wants does not solve your revenue problem.

00:14:12: Ouch!

00:14:13: But it's the truth and this kind of brings us to the harsh reality of marketing leadership.

00:14:17: today.

00:14:17: we've seen mechanics of visibility.

00:14:19: We've seen how top sales teams are encoding judgment.

00:14:22: but who is actually orchestrating all of this, like who is designing these systems.

00:14:26: Right now the data suggests almost no one.

00:14:29: really.

00:14:30: yeah that technology's moving at light speed but organizational readiness is lagging painfully behind.

00:14:35: Suzanne Pufard cited Gartner's twenty-twenty six CMO spend survey which found that seventy percent of CMOs want to be the primary leaders in AI within their organization.

00:14:44: sir

00:14:45: everyone wants to lead

00:14:46: But only thirty percent say they actually feel ready to scale it.

00:14:49: That

00:14:49: is a massive gap between ambition and capability, why are they so unready?

00:14:54: Well Liz Martin's insights give us the answer.

00:14:56: she reports that while seventy five percent of marketers have adopted AI in some form Eighty-four percent of them are still just using it to run generic, Run Of The Mill campaigns.

00:15:06: Just doing the same old stuff but slightly faster?

00:15:08: Exactly!

00:15:19: fragmented approval processes, and totally disconnected teams.

00:15:22: Yeah you simply cannot layer a hyper-efficient autonomous AI onto a dysfunctional bureaucratic org chart... ...and expect it to magically fix your company?

00:15:31: You really can't!

00:15:32: So what are the teams who're actually getting this right doing differently?

00:15:36: Sandeep Gulati says their very first move is avoiding the biggest trap in the market now which is over standardizing on one single tool.

00:15:44: Oh The One Size Fits All Fallacy, the idea that you just buy enterprise licenses for one platform and call it a day.

00:15:50: Exactly, you can't force one model to do every single task.

00:15:54: well.

00:15:55: Sandeep says the elite marketing teams map out their entire workflow first.

00:16:00: Workflow

00:16:01: First tools second?

00:16:02: Yes

00:16:02: only then did they assign specific models two specific jobs based on what the architecture of that model is inherently good at.

00:16:09: for instance You might use chat GPT for ideation and creative brainstorming because it's super conversational

00:16:15: right.

00:16:16: but Then you shift a quad deep analysis and long form reasoning because of its massive context window.

00:16:22: You pull in perplexity when you need real time research, and life citations And then use Gemini for seamless integration into your workspace documents.

00:16:30: That is true.

00:16:30: system design that's looking at AI as a composite architecture rather than just single software subscription perfectly aligns with where the industry is heading next.

00:16:41: Which is where?

00:16:41: Carolyn Healy shared Gartner projections, showing that sixty percent of brands will use agentic AI for one-to-one customer interactions by twenty-twenty eight.

00:16:50: Okay let's define agentic Ai for the marketer listening on their commute right now.

00:16:54: How was an agent different from what we're using today?

00:16:56: That's

00:16:57: a great question.

00:16:58: Today you prompt an ai and it gives you an answer.

00:17:02: then It stops

00:17:03: Right!

00:17:03: It waits For You.

00:17:04: An Agentic AI Is autonomous in goal seeking.

00:17:08: You give it an objective like, hey nurture this list of five hundred leads until they book a meeting and the agent figures out steps writes emails analyzes replies adjusted strategy and executes all tasks without you constantly prompting that.

00:17:21: That sounds incredibly powerful but also massive management nightmare if don't know what your doing

00:17:26: which is exactly problem.

00:17:27: René Jones highlighted BCG data showing mass disconnect here.

00:17:32: Ninety-six percent of CMOs claim AI is driving end to end transformation in their companies.

00:17:38: Ninety six percent?

00:17:39: Yeah, but yet only eight percent are actually running multi aging campaigns.

00:17:43: The reality is the CMO role has to fundamentally shift.

00:17:46: you can no longer be a campaign architect.

00:17:49: You have to become a system designer.

00:17:51: You aren't designing the creative anymore.

00:17:53: your designing the parameters and ethical boundaries for machine that generates the creative.

00:17:58: And while leaders are trying to figure out how to design these boundaries, there is this massive ticking clock on the regulatory side that almost no one is talking about.

00:18:06: Of EU regulations?

00:18:07: Yes!

00:18:08: Felix Schlager flagged at the EU AI Act's transparency and labeling rules are enforceable right now not in twenty-twenty seven...now

00:18:16: That has huge implications for deployment.

00:18:19: what exactly does that mean for say an AISDR?

00:18:22: It means Any AI system that interacts with a human, whether it's the customer support chatbot on your website or an AISDR sending cold outbound emails must explicitly disclose that is an AI.

00:18:35: Wow!

00:18:35: Yeah If you try to pass agent off as a human SDR named Dave who just loves grabbing virtual coffee You are in direct violation of law.

00:18:44: I can imagine slapping and iamabot disclaimer on an outbound email will severely impact conversion rates.

00:18:50: But what happens if you just ignore it?

00:18:52: The

00:18:52: fines are staggering up to fifteen million euros or three percent of your global turnover for non-compliance.

00:18:58: Fifteen million euros, that changes the calculus entirely!

00:19:02: You can't Unleash these goal-seeking agents into the wild without strict governance and compliance frameworks built in to the core system from day one.

00:19:11: Exactly, it's like we're handing out Formula One cars to marketing teams but the track they are driving on is made of dirt.

00:19:18: They don't have a steering wheel And they don't even know the rules of the road.

00:19:21: That's a great analogy.

00:19:22: If you don't design this system Establish the governance and prep the organization The car just crashes faster.

00:19:32: The technology has vastly outpaced the infrastructure to support it, that's for sure.

00:19:37: But if we connect all of this—the shift in visibility... ...the encoding of sales judgment….

00:19:41: …the rise of economists' agents.... I want leave you with one final mind-bending paradigm shift from Amel Asin Padilla.

00:19:49: Amel argues that we are rapidly moving beyond BDB business to business and entering the era of B to A. Business to agent?

00:19:58: Yes!

00:19:59: Agent mediated purchases are projected to top fifteen trillion dollars by twenty-twenty eight.

00:20:04: Fifteen trillion dollars, just let that number sink in for a second!

00:20:08: Think about what this means your strategy tomorrow morning.

00:20:11: Your brand is no longer competing with human attention.

00:20:14: A human buyer might not even do the initial vendor research anymore.

00:20:18: Right You're brand is competing to be selected by logical highly efficient completely unsentimental AI agents That are tasked finding best software for their company.

00:20:29: So if your digital presence isn't structured to be read, interpreted and fully trusted by a machine...

00:20:35: You will be entirely excluded from the consideration set before human buyer ever even sees you name.

00:20:41: The agent will compile short list of say three vendors for the Human to Review And If AI doesn't understand co-occurrence Your authority and consistency simply won't be on that list.

00:20:52: It is ultimate invisible filter

00:20:54: That completely redefines what it means to build a brand in this decade.

00:20:58: You aren't just pitching people anymore, you are pitching the algorithm that works for them!

00:21:03: If you enjoyed this episode new episodes drop every two weeks.

00:21:07: also check out our other editions on field marketing channel and partner marketing account based marketing martech go-to market social selling.

00:21:16: Thank you so much for joining us on this deep dive.

00:21:18: Take a hard look at your brand's digital footprint tomorrow because remember, well machines are reading.

00:21:23: make sure giving them good story.

00:21:25: hit that subscribe button and we'll see ya next time.