Best of LinkedIn: AI in B2B Marketing CW 29/ 30

This episode is provided by Thomas Allgeier and Frennis, based on the most relevant LinkedIn posts about AI in B-to-B marketing in calendar weeks twenty nine and thirty. 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 AI. And for those of you listening who are strategic B-To-B Marketing professionals today Is really about getting under the hood. if what's actually working right now.

Yeah, exactly. because I mean imagine a key enterprise buyer is researching software for their company. They don't go to Google anymore... Right they just don't!

they ask ChatGPT to compare you and your biggest rival And in about two seconds the AI doesn't list your features. No it invents massive downside of your product. Yes based on like a two-year old Reddit thread The buyer crosses off the shortlist immediately and you never even knew they were looking, which is terrifying. It's the reality of B to D marketing right now though I mean if you look at the landscape there is just so much noise.

So our mission today is cut through that hype. Yeah we have to look at actual operational shifts happening on the ground. Right Because finally moving past the era of simple prompt hacks If your are trying to figure out how write a slightly better prompt for blog post. you're missing the structural revolution.

You are completely missing it. Today is about systems that B to be leaders building in this agentic era. It's about rewiring your entire revenue engine, basically. So let start at very beginning of that Revenue Engine Because before you can sell anything have actually been found.

Right Discovery Exactly. But Discovery has fundamentally moved off the traditional brand website, it's gone straight into AI answer engine. It is a complete shift in like physics of buyer behavior Generative Engine Optimization or GEO and Answer Engine Optimisation AEO. Those are basically replacing traditional SEO now Totally And to really understand what that looks for your pipeline Sara Soleimani pointed out reality I think should wake up every go-to market leader.

Oh The shortlist thing. Yes She noted that B-to-B buyers are now forming their vendor shortlists entirely inside AI tools. Before they ever even visit a brand's website? Exactly!

By the time you hit your landing page, The decision is... well it basically already made And That creates massive problem for marketers Because the rules of visibility Are now completely fragmented. How so? Like compared to old days.

Well think about traditional SEO. Right You optimize Google and that basically covered your bases for everyone. Right Google was the only game in town, but Andrew Warden shared some really eye-opening data on this. he compared google AI overviews to Gemini.

okay no Google controls both of these, obviously. Yet there is only a thirty percent overlap between sighted sources and mentions on Gemini compared to a sixty-four percent overlapped on Google AI overviews. Wait really? Yeah So even within the exact same google ecosystem The AI's looking at entirely different signals To decide if your brand Is worth mentioning.

Exactly Shared corporate infrastructure does not mean shared rules anymore And it gets vastly more complicated across the different platforms. I can imagine. Rochelle Tognetti broke down how each major AI evaluates trust, actually. She noted that ChatGPT heavily rewards credibility established on places like Wikipedia and Reddit...

Okay! Wikipedia and Reddit for chat GPT Right But Gemini on the other hand tracks your presence at Google Shopping. Google shopping? That's so different I know.

And then if you want to show up in Google AI overviews The model leans heavily on YouTube content. Okay, let's unpack this metaphorically because it's wild. It's like taking the bar exam in fifty different states. Yeah that is a good way to look at it.

You can't just study law broadly anymore. you have know what specifically matters to judge in specific room. Exactly Chachi BT is a Judge who loves Wikipedia and Reddit debates. Gemini is a judge only cares about your retail footprint.

So how can a brand possibly build authority across all these, you know fragmented contradictory rules? You have to focus on something called the Citation Core. The Citation core... Yeah Alexandra London analyzed Shopify's visibility in chat GPT To figure out why they completely dominate.

And Shopify presence isn't driven by its own massive blog. Oh it's not. No their visibility is anchored By over a thousand citations from G-II and like eight hundred citations From Kaptera Wow. It perfectly aligns with what Thomas Ross outlines in his Vantage scorecard, actually AI models borrow trust from third parties.

right they do not trust your own marketing copy. okay let me make sure i'm getting the mechanics of this. if I publish say a thousand amazing blog posts about why my software is the absolute best The AI essentially discounts that because im biased. But if a hundred independent software review sites say I'm the best, the AI takes that as like mathematical proof.

Precisely it's a probabilistic model. The AI triangulates your brand value based on what the aggregate web says about you. so You can't buy your way in nope? You cannot pay your way into the sources.

AI trusts you have to earn That density of mentions elsewhere which makes the way these engines display results Incredibly high-stage. Oh for sure. Aruprasad ran a brilliant test on this exact dynamic. He created five-axis tests to evaluate AI answers.

Right, looking at things like presence slot? Yeah! Presence Slot Coverage Liability and Door. But I really want zero in that liability axis because when i read it's stopped me in my track.

It is! AI platforms will now assign unprompted objections to your brand before the buyer even speaks. Exactly, it fundamentally changes sales. It really does!

Imagine a prospect asks an AI to compare you and the competitor? The ad doesn't just give a neutral list of features anymore…. Right...it's not just bulleted lists..

It actively generates a downside or liability for your brand may be citing complex implementation process or lack customer support And bases this on aggregated internet sentiment Exactly....and presents that liability as objective fact. That is wild because objection handling used to happen on a live sales call, right? Where a rep could actually provide context or maybe on a carefully crafted landing page.

Right you had control but now the AI handles the objection for the buyer in a split second completely out of your control. You literally can't defend yourself No! So how do we even start measuring this if it's all happening inside private chat window. Well were just getting visibility tools.

Nirapunj highlighted that Google Search Console is finally offering first-party reporting on AI overviews and AI mode impressions. Oh, really? That's huge! Yeah it's basically the first real window marketers have into organic AI visibility.

And on the paid side things are moving just as fast right. Bill Stathopoulos shared that ChatGPT just opened self-serve ads. Yes, at a twenty five dollar daily minimum. I think Yeah but the wild part is there's zero keyword targeting.

Not all You describe who you want to reach and AI matches your ad directly into relevant user conversations dynamically. That completely shifts the entire paradigm. But let us follow the buyer journey here for second. Let say you've successfully gamed the answer engine You built your citation core, you're on the AI shortlist and The buyer is now fully aware of you.

Right? Awareness doesn't close deals. Yeah How do you actually capture that intent? That's where the traditional outbound stack Is completely falling apart right Now.

yeah totally because if you are highly visible to Ai but Your Outbound motion is still running On some manual fragmented You know six toolstack from twenty nineteen You can't capitalize on that demand And we Are seeing a complete overhaul of how outbound is structured to fix that exact issue. Shabir Mohsen shared a perfect operational example with this, he's using an AI go-to market agent that monitors over forty different buying signals simultaneously. Forty signals? Yeah it completely replaces the manual outgun stack.

It researches the prospects cross references all forty signals drafts sequences and runs entire workflow. So humans just step in for actual sales calls Exactly And look at scale. when deployed correctly Alex Vaca detailed a workflow using just four prompts for a sales development representative. Just four prompts?

And it was scaled across more than two hundred and seventy-five B to B companies. Those four core prompts generated seven point eight three million dollars in pipeline. That's unbelievable! It proves that deep research and list building are now fundamentally agent work.

Oh, and speaking of scaling that research there was the pre-event scoring model too. Oh right! Vladimir Blagojevich's Model Exactly. He took thirteen hundred event attendees And automatically triaged them down to a hundred nineteen high value targets Before even happen Right before any conversations take place.

Wait hold on though I have push back here A little bit. Ok go for it. Forty signals or Scaling prompts across hundreds companies. Most sales teams I know can barely handle basic intent data from their own website without drowning in false positives.

That's true, and then they just end up spanning people. So how is a hyper automated AI agent not just creating more noise? I mean are we just automating the wrong part of the funnel here? since it valid question because Steve Armenti pointed out that buyers initiate eighty percent of first conversations themselves.

Deals are ultimately decided by familiarity and relationships, so why aren't we trying to speed up cold-outbound if the buyer has already made their shortlist through AI? I love you brought this up! That is exactly what most companies fall into right now... Automating a broken process just means you fail at scale.

Exactly, faster garbage? basically Yes and your pushback raises the most critical issue in AI. go to market right now which is the underlying data bottleneck. The data decay.

Right. Ryan Staley observed that B-to-B data decays at roughly thirty percent of year. Wow! Thirty percent.

Yeah People change jobs. Companies pivot Budgets freeze. This datarot Is exactly why ninety five percent Of enterprise AI pilots fail. So it's not that the AI model itself isn't smart enough, It is we are feeding in absolute garbage.

Precisely! If you feed an AI stale wrong or thin data A highly confident perfectly written email sent to a person Is actually worse than doing nothing at all. Damage as the brand. Exactly You don't need smarter models You need honest pristine data.

Majvoji's case study on owner.com illustrates this perfectly. Oh, I saw that they hit nearly a hundred million dollars in annual recurring revenue right? In under four years.

But they didn't do it by just turning on quote-unquote AISDRs to blast more emails. They did it by managing system throughput and driving deep organizational change around their data. They treated AI as an operational shift, not just a software update. Which is such a crucial distinction.

and actually speaking of operational shifts if you listening are finding these structural changes in B-to-B marketing As fascinating as we are? make sure to hit subscribe so you catch all our future deep dives. We are tracking this evolution constantly because that organizational change Is really where the rubber meets The road exactly. You clearly can't run an AI native outbound motion with a siloed?

Structure or marketing teams has to fundamentally break down and rebuild. It has too. If you are a CMO listening right now, just look at your org chart. You probably have team of channel specialists an SEO person A paid social person An email person And according to Carolyn Healy who referenced Gartner's twenty-twenty eight projections That structure is going be completely obsolete Really?

Yeah! Gartners states that sixty percent brands will use agentic. one the one adoption. This effectively spells the end of traditional channel-based marketing structures.

Wow, so The Marketing Leader's job is completely shifting? It's shifting from being a campaign architect to be an systems designer. But restructuring team is intensely political. I mean people are protective of their roles.

Oh absolutely John Miller shared great story about DAXCO. that shows what it actually takes to do this. DAXCOS currently running over Eighty A.I agents, eighty.

but they didn't get there by making it a side hustle. They didn't just tell their team to you know play around with chat GPT on Friday right? It wasn't a hobby No! They only achieved that scale By pulling the vice president entirely out of their day job To build a dedicated AI center of excellence.

Taking a VP offline to focus solely on AI orchestration is a massive commitment. It really is! But it's necessary because, as Chris Culler pointed out, marketing roles are shifting back toward generalists rather than narrow specialists. Interesting.

why generalist? Because the AI is the ultimate specialist now...it can write code and draft copy and analyze spreadsheet which you need. our generalists who can orchestrate all those special agents.

But let's talk about the danger of getting that orchestration wrong, because right now for The Average Team it feels like an orchestra where every single musician is playing a completely different song in their own practice room. That has very accurate picture. Right. one person uses Claude another use as custom GPT someone else using AI writing tool.

they all have different prompts and absolutely no shared standard And Orin Greenberg highlighted this fragmented individual. AI adoption is actually the single largest tax on BWB marketing ROI right now. A tax on ROI? That makes so much sense!

It creates this illusion of productivity. when everyone uses AI for low friction tasks in their own silo you don't get a symphony, You just get a racket Right. The team was operating at exactly same headcount with slightly better grammar. To fix this you have to prove the fundamentals first.

Okay, but how do you do that without just adding more tools to the pile? You'd do it by hand! Andrei Zenkovich broke down on how Backbase handled this transition. They ran a completely manual account-based marketing pilot first...

to prove their go-to market fundamentals Manual like no AI at all None. they figured out the exact sequence. The messaging got real human to reply All manually. They didn't let AI automate anything until they had proven the process actually worked.

Because you cannot automate a process your team hasn't nailed down manually? Exactly! If the manual process is broken, AI just scales the chaos And that brings up the risk of teams trying to build their own workarounds. Rachel Roundy talks about the Ikea effect in software.

Oh...the Ikea Effect? Yeah it's the danger of teams vibe coding there own messy fragile tools instead relying on stable enterprise grade infrastructure. Right, because a marketer might build the clever prototype that looks great in a demo but it completely falls apart production when hits real customer data.

Exactly! Alex as a party's update on service now reinforces this reality. actually they saw a nine-fold jump in production agentic AI customers The Nine Fold Jump. Yeah But that massive scale was driven entirely by strict governance not just deployment speed.

Safe orchestration at scale is where the real work is because governance sounds super boring until your vibe coded AI agent starts hallucinating custom pricing to a key enterprise account and legally binds you too, right? Precisely. Governance Is The Breaks That Allow You To Drive Fast Without Crashing Your Brand. Okay so let's bring this high-level org transformation down to the individual practitioner.

if we have this new team of BtoB generalists how do they actually work with AI day today it's changing fast Because we are definitively moving away from just typing a question into Every single day at seven AM. a workflow runs automatically. It scans competitor updates, pulls account intelligence from across the web and drops a synthesized summary for him before he even wakes up. No code required no manual trigger Exactly.

And Zohi Mustafa expanded on this with the concept of reusable Claude skills. Instead of a marketer writing one off prompt over-and-over again to say analyze a campaign They package that capability into a repeatable building block. But the real magic of those building blocks is what they can connect to, right? Yes and that's where Model Context Protocol or MCP comes in.

Can you explain MCP really quickly? for who might not know? Sure Think of MCP as this secure universal translator between your AI and actual databases. So instead if downloading a spreadsheet from tool and pasting it onto Claude MCP allows Claude securely reach through that tool run query itself And pull exactly what he needs.

That is a massive workflow shift. I mean, Anna York used an MCP connection with Semrush and it completely changed her output. Oh yeah! Her keyword research?

Yeah she cut our keyword research time from four hours down to twenty minutes but It wasn't just about doing the exact same task faster. The MCP allowed her to shift her entire focus Right From keywords to questions. Exactly. Instead of traditional SEO asking what keywords people search for, the MCP allowed her to focus on specific questions that AI needs to answer before it decides to cite a brand.

It's like moving from checking ten different disconnected dashboards... to just asking a hyper-competent colleague who has already memorized every single file in your server and can synthesize instantly. Exactly! Amelia Mahler made a fascinating point along these lines too.

She talked about connecting live meeting transcripts and client briefs directly inside Claude. So the AI isn't giving generic internet scraped advice? Right, it's looking at Live Proprietary Data And Giving a Prioritized Action Plan Based on that. Which really makes me wonder Does this mean Context Engineering is now officially more important than Prompt Engineering?

Absolutely The prompt Is just trigger or context is intelligence. That's great way to put. Kieran Flanagan confirmed this dynamic perfectly. He noted that a marketer's ultimate leverage used to be domain authority, right?

If you had high-domain authority your content won on Google. Today AI and automated workflows are the new leverage. The variable that separates a brilliant output from a useless one using the exact same AI model isn't how cleverly you phrased the prompt. It's the context.

It is the depth, quality and structure of the contexts you provided. Context wins every single time Even in highly creative work. Jessica Aries tested a carousel design across four different AI tools. Oh I saw this test.

Yeah, she found that tools like Claude Design actually make structural design decisions based on the context provided rather than just blindly filling in a visual template. So it actually understands what its making? It understood the hierarchy of information and arranged the design logically much like human designer would purely because the context was engineered correctly. Wow, so we've traced this incredible evolution today.

We've gone from the black box of discovery in AI answer engines to restructuring how we capture intent through AI driven outbound. We looked at why you have to tear down your old marketing org chart, to govern all this. Right! All the way down into daily context engineering of an individual marketer's workflow.

it is a complete structural rewiring of B-to-B Marketing. It Is. And if we pull these threads together there s one final profound implication that every go-to market leader needs to internalize today. Okay What s That?

Well Patrick Berghoff introduced his concept of the Oscillation Tax. The Oscillations Tax. what does actually mean in practice? Well, if we connect the oscillation tax back to our own Prasad's liability axis from our very first theme.

Oh right! The idea that AI generates unprompted objections? Exactly...the warning here is stark If you cut your customer experience or brand building budget today just save money for a quarter.

You aren't taking temporary dip in metrics. Okay You are permanently cementing negative reputation into weights of an LLM. Wow Let that sink in for a second. You are literally baking your own liability into the machine exactly because AI models train on the aggregate sentiment of the web over time if Your service slips or your brand goes quiet and stops earning.

those third-party citations The negative reviews, and the lack of validation become part of the model's institutional intelligence And you can't just buy a top Google ad to fix it next year. no The machine has a very long memory. Your brand's institutional intelligence and relationship capital are being judged by algorithms right now. So if you're invisible today?

Or worse, If your viewed as the liability of AI today You simply won't even make it shortlist tomorrow. That is a chilling thought honestly but incredibly necessary to understand! You are building equity with this machine every single day And just can't fake that... You really cant If you enjoyed this episode.

new episodes drop every two weeks. Also, check out our other editions on field marketing channel and partner marketing account based marketing. Martek go to market and social selling. Thank you so much for joining us for this deep dive And don't forget to subscribe So you don't miss out on what's next.

we'll see you next time.