Best of LinkedIn: AI in B2B Marketing CW 07/ 08

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

00:00:06: In calendar weeks seven and eight.

00:00:08: Frenness is a B-to-B market research company helping enterprise marketing teams sharpen their strategies and outreach with customer segmentation ideal customers profiles deep dives customer needs analysis buying center insights.

00:00:22: Welcome to the deep dive, everyone.

00:00:24: We are really getting into a massive shift today.

00:00:26: look if you've been following The Noise on LinkedIn lately yeah You probably know that AI in B-to-B marketing has It's officially moved way past That shiny new toy phase.

00:00:36: Yeah it really feels like the party is over and the cleanup crew Has arrived but ya'know In A good Way!

00:00:42: That whole hangover phase of just playing around with basic chatbots Its done.

00:00:46: Looking at the sources we have from week seven & eight The vibe has shifted completely.

00:00:50: Right, we aren't talking about generating funny images or turning out generic blog posts anymore.

00:00:54: the sources are painting a much more serious picture now.

00:00:58: AI is actually becoming the operating system for sales and go-to market.

00:01:02: Exactly it's moving from a tool you use sometimes to the actual infrastructure You live in always yeah And that transition changes literally everything About how your company is going to operate.

00:01:11: It's getting real and its getting measurable

00:01:13: It is.

00:01:14: So to help you wrap your head around this, we've clustered the insights from these two weeks into three main areas.

00:01:20: First We're going to dissect the evolution of AI sales teams.

00:01:25: Specifically This move From basic bots To what they are calling orchestrated agent swarms

00:01:30: Which I know sounds like a sci-fi horror movie But as will see it's really just efficient economics.

00:01:36: Yeah

00:01:36: totally.

00:01:37: Then, we are heading into the boardroom.

00:01:40: We're seeing a hard shift in go-to market strategy from experimentation to strict P and L accountability.

00:01:46: basically if it doesn't make dollars does not makes sense anymore.

00:01:49: The free ride is over Yep.

00:01:50: And finally will unpack new visibility Will explain why SEO is evolving into GEO generative engine optimization Why your humble about page might actually be single most critical asset for training AI models that customers use right now.

00:02:05: So let's dive right into that first theme, AI sales teams and the infrastructure backing them up.

00:02:11: There is a full-blown revolution happening here, but it's fascinating because It's coming from two completely opposite directions.

00:02:17: On one end you have this DIY disruption.

00:02:20: I was reading a post from Ali Juani that honestly made me question the entire sauce pricing model.

00:02:25: Oh!

00:02:26: You're talking about The twenty five dollar agent?

00:02:27: Yes

00:02:28: So Ali Juane shared That he built an AI sales assistant using An open source framework called Open Claw.

00:02:34: And when i say Assistant...I don't mean A little widget in the corner of your screen.

00:02:37: He claims This thing handles lead research writes personalized emails, monitors the inbox handles follow-ups.

00:02:43: The whole nine yards!

00:02:45: And a cost was the real shocker there.

00:02:47: Roughly twenty five dollars a month in compute credits?

00:02:50: He's actively comparing this to enterprise tools that charge two thousand dollars per month per seat.

00:02:56: His argument is most of these expensive sauce tools are essentially just wrappers around large language models.

00:03:01: they have nice user interface

00:03:02: Right.

00:03:02: he basically asking why pay rent when you can build house for the costs?

00:03:08: Exactly.

00:03:09: It's a really compelling argument for the democratization of tech, it forces us to ask what we are actually paying for.

00:03:16: but while the DIY approach is fascinating for saving a few bucks The real enterprise value at least from other sources that were reviewed seems to be coming from architectural specialization.

00:03:27: You're referring to the SAM system that Thorsten L discussed

00:03:30: Precisely!

00:03:31: ThorstenL shared breakdowns with a system called SAM.

00:03:34: The key differentiator here is that SAM isn't an AI.

00:03:37: It's not one overworked robot trying to do everything at once.

00:04:07: Another is responsible only for enriching prospect data.

00:04:10: A third writes the messages, a fourth books the

00:04:13: meetings.".

00:04:13: So it's essentially a digital assembly line?

00:04:16: It's an orchestrated swarm and Thorsten noted something really crucial about it.

00:04:21: they aren't using generic prompts.

00:04:23: They utilize twenty-seven specific industry playbooks.

00:04:27: The way Sam sells to healthcare procurement Is fundamentally different from how it sells to a SaaS founder.

00:04:33: That context is exactly what's been missing.

00:04:35: I think we are all so tired of the spray and pray AI emails that sound like they were written by a robot, but it never actually met human beings... It's generic-I hope this email finds you well-nonsense!

00:04:47: But that brings us to the unsexy hero of this whole story which Spencer Perique highlighted brilliantly.

00:04:53: You can have the best agents in the world.

00:04:54: You could have Sam or Ali Jihwani open clause set up.

00:04:57: but perique argues your AI isn't broken Your infrastructure.

00:05:01: That's a bold take.

00:05:02: What does he mean by infrastructure in this context?

00:05:04: Is he just talking about having a clean CRM?

00:05:07: It goes way deeper than just CRM hygiene.

00:05:10: Perique's team focuses on defining enrichment and scoring rules before a single email ever goes out.

00:05:16: He claims that, by filtering rigorously up front they actually remove forty to sixty percent of the waste.

00:05:22: So they aren't using AI to spam more people.

00:05:25: They're using logic to spam fewer people.

00:05:27: Exactly it is all about subtraction.

00:05:29: But really smart part is moving away from those calendar-based trips.

00:05:34: You know the standard drill?

00:05:36: Email one goes out on Tuesday, email two goes out Thursday...

00:05:39: Which I automatically archive without even reading!

00:05:42: We all do.

00:05:43: PARIC advocates for behavior based triggers.

00:05:46: If a prospect visits specific case study at your site they get completely different flow than if you just open pricing page.

00:05:53: It's about reacting to intense signals in real time not filling up schedule.

00:05:58: Speaking of intent and infrastructure, we have to talk about the Salesforce case study from Sastro.

00:06:04: Jonathan MK and Jason and Lemkin shared this in.

00:06:06: The Numbers are just well.

00:06:07: they make you look twice!

00:06:08: This is the ShelfWare software story right?

00:06:10: It

00:06:10: IS!

00:06:11: Jason Lemkin admitted that for a long time sales force had become shelfware for them.

00:06:15: They were barely using.

00:06:16: it was an expensive database gathering dust But then turned into central nervous system For over twenty AI agents Using something called Agent Force.

00:06:28: They finally connected the brain to the limbs.

00:06:30: What was this specific application they used it for?

00:06:32: They went after ghosted leads.

00:06:35: These were a thousand warm inbound Leads that human reps had just dropped no follow-up lost on The Shuffle which

00:06:43: let's be honest happens and literally every sales organization Leads slip through the cracks because humans get busy or distracted.

00:06:50: right,

00:06:51: but here is the stat that blew my mind.

00:06:53: Agent Force achieved a seventy-two percent open rate on those emails.

00:06:57: Hold on, seventy two percent?

00:06:58: Yeah The industry average for cold outreach is lucky to hit two to four percent.

00:07:03: That's the power of context because it wasn't cold!

00:07:06: The AI

00:07:07: had full history, passive end attendance Sponsorship details Every previous interaction.

00:07:12: It was just hey buy our stuff Hey welcome back here something relevant based what you did last time.

00:07:17: that

00:07:17: really underscores why this CRM Is the database record.

00:07:21: If that data hadn't been there, even if the humans weren't actively using it.

00:07:25: The AI would have been completely blind.

00:07:27: It proves that data hoarding if you structure it correctly actually pays off eventually?

00:07:32: It does.

00:07:33: but before we declare victory and fire all the sales reps We do need a reality check here.

00:07:38: Spencer Lyna shared some data on a company called Monaco Which is backed by Founders Fun And it highlights A pretty dangerous trap!

00:07:45: The conversion trap.

00:07:46: Yes

00:07:47: Pure AI drives massive volume at the top of the funnel.

00:07:50: You get more leads, you get faster responses but Luna points out that when you get down to complex deals pure AI totally fails to convert!

00:07:58: You simply cannot automate trust on a six-figure deal.

00:08:01: So what's this solution then?

00:08:02: Do you just turn the AI off at that stage?

00:08:04: No...you go hybrid.

00:08:06: The data showed that hybrid teams where the AI handles the prospecting and sequencing But humans step in to handle closing.

00:08:14: They see thirty percent shorter sales cycles and get this, twenty percent larger deal sizes.

00:08:20: That makes intuitive sense.

00:08:22: you use the machine for the leverage in the volume And You Use The Human For The Nuance In The Trust.

00:08:28: This actually segues perfectly into our second theme today.

00:08:31: Go To Market Strategy Because If We Are Changing How We Sell Using These Swarms & Hybrid Models we have to fundamentally change how measure success

00:08:40: And the boardroom is apparently getting very impatient.

00:08:43: Cassie Young issued a pretty stern warning to CROs recently, she says boards are completely done rewarding AI experimentation.

00:08:52: The era of hey look at this cool pilot we ran...is over.

00:08:55: It's all

00:08:55: about P&L impact.

00:08:56: now

00:08:56: Exactly She introduces this prime framework?

00:09:00: The specific metrics are deep but core questions are brutal.

00:09:04: Is your revenue per head actually climbing?

00:09:07: And is your CAC, customer acquisition cost payback period decreasing?

00:09:12: That revenue per head metric.

00:09:15: Think about it.

00:09:16: if you implement all this AI and your revenue per head stays flat You haven't actually increased productivity.

00:09:21: You've just added software costs, made the entire operation less profitable.

00:09:25: Right!

00:09:25: If you're writing bad emails faster but not selling more efficiently The board does NOT care.

00:09:30: But some teams are nailing that transition by looking at problem differently.

00:09:34: I loved the story from HockeyStacks shared with Kyle Poyar and Amir Ali.

00:09:38: They didn't plug in a bot to write e-mails.

00:09:41: they reverse engineered their best humans.

00:09:43: This is process mining essentially Cloning the successful behavior, not just the output.

00:09:48: Yes!

00:09:49: They looked at their top account executives and asked what are they doing that average reps aren't doing?

00:09:55: And found out that the top reps were doing these really high-touch things sending physical gifts creating custom Slack channels for the prospects Which

00:10:03: usually scales terribly.

00:10:04: You can ask every single rep to manage fifty custom Slack Channels

00:10:09: perfectly Normally.

00:10:10: no you cant.

00:10:11: but they built next best action agents.

00:10:14: These agents didn't do the work for the rep, they prompted the rep.

00:10:18: The AI would nudge them and say hey this prospect just hit the stage.

00:10:21: send a gift right now or it's time to open the Slack channel.

00:10:25: It is like GPS for sales process but it's telling you exactly when to turn so you don't miss the exit.

00:10:32: And,

00:10:32: it worked beautifully!

00:10:33: Win rates went up five-to seven percentage points and cycle times dropped by nearly twenty percent.

00:10:39: It wasn't about replacing The Human who was making every human perform like the top one percent.

00:10:44: That aligns perfectly with a great quote from Matt Hines regarding orchestration.

00:10:48: He said AI doesn't create chaos...it exposes the chaos that is already there.

00:10:56: Seriously, think about it.

00:10:57: If your messaging is inconsistent AI just helps you produce inconsistent messaging much faster.

00:11:03: Orchestration is non-negotiable.

00:11:05: now You can't have rogue agents running around making up their own rules

00:11:09: Speaking of rogue agents and changing the rules.

00:11:11: Jonathan Schroyer had a pretty radical take on all this.

00:11:14: He thinks that traditional linear sales funnel was completely dead.

00:11:19: A lot people say so but Schroier actually offers replacement Learning loops.

00:11:23: Right The idea is that instead of a straight line from awareness down to purchase, you have AI continuously testing and refining based on profitability.

00:11:33: It's continuous loop of optimization.

00:11:34: Its

00:11:35: dynamic versus static...a traditional funnel as aesthetic map A learning loop like ways it re-routing in real time based traffic conditions what actually working?

00:11:44: And the dynamic nature disrupting more than just sales funnel completely upending how companies get found first place.

00:11:52: Which brings us to our third theme AI search and visibility.

00:11:57: This is the stuff that should really keep marketers up at night.

00:12:00: The acronyms are changing, And it's not just semantics.

00:12:04: Christon Lee & Leah Gabriel Narik broke this down.

00:12:06: We all know SEO Search Engine Optimization That's Just.

00:12:11: Please Google put me on page

00:12:13: one.

00:12:13: But now we have AEO Answer Engine Optimization.

00:12:16: That's where you want to be the direct answer when someone asks a question, like a featured snippet but on steroids.

00:12:21: And then the new big one GEO Generative Engine Optimisation.

00:12:25: This is the holy grail right now.

00:12:27: Uh-huh, this about being cited by chat gpt or perplexity or clawed when a user explicitly asks for recommendation.

00:12:34: It's not about providing a blue link anymore.

00:12:36: it's about being baked into the synthesized answer itself.

00:12:39: And this isn't just for consumer questions like what's the best pizza near me?

00:12:43: Vinyamin made a point that I think A lot of b to be marketers are completely missing Right Now.

00:12:48: he said technical buyers aren't googling any more.

00:12:50: no they really aren't their coding.

00:12:53: He mentioned that developers are asking tools like Cursor or Claude Code for API recommendations while they're actively inside their integrated development environment.

00:13:01: That is wild to think about.

00:13:03: So if you're a marketer optimizing your blog for Google keywords, but the developer you want to reach is asking their coding assistant for recommendation right inside there code editor?

00:13:13: You are effectively invisible to them!

00:13:15: You

00:13:15: literally don't exist in their world.

00:13:18: If your brand isn't visible inside that IDE workflow... ...you were entirely out of game.

00:13:23: The big question was how do we get recommended?

00:13:25: How do you hack this new system?

00:13:27: Jeff Gibbons has some really sobering stats on this.

00:13:29: He studied over six hundred and fifty brands, And he found that the average brand is mentioned about eighty percent of time in AI answers.

00:13:38: but...and then it's a kicker.

00:13:40: It is recommended less than twenty-five percent at times.

00:13:42: That Is such a crucial distinction.

00:13:44: Being mentioned as just an AI saying I know this company exists.

00:13:48: being recommended As the AI saying i actually trust This Company to solve your problem.

00:13:53: Gibbons points out that AI models are inherently risk-averse.

00:13:57: To get a real recommendation, you need what he calls convergent signals.

00:14:02: You know what?

00:14:13: Exactly!

00:14:22: it needs external corroboration.

00:14:24: But speaking of your own website, there's one thing you can actually control right now.

00:14:28: Anna York gave a very specific tactical tip that everyone listening should probably go do today.

00:14:34: Ah yes the about page optimization.

00:14:36: Yes

00:14:36: She says your about page is the gold standard for AI training.

00:14:39: If an AI spider crawls your site to figure out who you are, that's very first place it

00:14:43: goes.".

00:14:44: She suggests optimizing with a question-based title tag literally naming the page Who Is and then having a clear fourty to sixty word intro summary right at top

00:14:54: And she heavily emphasized structured schema data too.

00:14:57: Can

00:14:58: explain this for non technical listeners?

00:14:59: Why does Schema matter so much for AI?

00:15:01: Think of Schema like labels in pantry.

00:15:04: Without it, the AI just sees a glass jar of red stuff.

00:15:07: It could be salsa or jam and paint.

00:15:10: Schema is the physical label that says this strawberry jam.

00:15:14: In web terms tells machine.

00:15:15: This specific text is CEO's name The product price.

00:15:20: This text is our headquarters location.

00:15:22: If you make it mathematically easy for the machine to understand You it statistically more likely.

00:15:27: explain correctly.

00:15:29: It seems so simple, yet I bet ninety percent about pages out there are just fluffy mission statements about changing the world or synergizing paradigms.

00:15:36: Exactly!

00:15:37: The AI does not care about your fluff.

00:15:39: it wants facts entities and relationships.

00:15:42: give us data.

00:15:43: Let's move to our final cluster here tools and human moat.

00:15:47: because with all this automation agents learning loops generative search that existential question becomes what is actually left for humans do?

00:15:56: Kaiam Calvert had a take that I found really reassuring.

00:15:59: The human edge!

00:16:00: He argues, as AI content scales to infinity the value of generic content goes straight to zero if everyone can generate an ultimate guide to B-to-B sales in five seconds.

00:16:11: nobody cares about your guide anymore.

00:16:13: it's just basic supply and demand.

00:16:15: infinite supply reduces the value at commodity near zero.

00:16:19: So authentic founder-led content becomes the only real moat you have left.

00:16:24: And Alex Olly backed this up with a really funny observation, he noted that physical gifting and swag are thriving right now!

00:16:31: It completely makes sense in a world of digital slop, as he calls it.

00:16:35: Receiving a physical object to hoodie?

00:16:37: A handwritten note feels incredibly rare.

00:16:39: It creates this biochemical reaction that a digital token or an automated email just can't replicate.

00:16:44: It physically proves that a human was

00:16:46: involved.".

00:16:46: AI Slop is definitely going into my daily vocabulary.

00:16:49: but you know AI isn't creating slopp.

00:16:51: its also coaching us!

00:16:53: Ryan Staley pointed out the amazing feature and clawed code called slash insights.

00:16:57: This is fascinating.

00:16:58: It actually analyzes your past sessions, how you've been interacting with the AI over time and it tells you how to prompt it better!

00:17:05: Its essentially the AI coaching human to be a more efficient operator.

00:17:09: That's so meta The tool is actively teaching you How To Use The Tool Better.

00:17:14: And finally A quick but really vital tip from Jennifer Allen regarding her tech stack.

00:17:19: She strongly advises using paid versions of tools like Claude or Proplexity

00:17:24: Because Of Data Privacy.

00:17:25: Exactly If you're feeding proprietary go-to market strategy or sensitive customer data into a free model, You might inadvertently be training the public model with your secrets.

00:17:35: Paid values usually have strict data privacy clauses.

00:17:38: It's very small cost for protecting intellectual property.

00:17:41: Okay, let's unpack this whole deep dive because we have covered a ton of ground today.

00:17:45: We started with the idea that AI is now at the operating system not just a side tool.

00:18:00: And we saw that the set-it and forget it mentality is dying in The Boardroom.

00:18:04: You need P&L impact, you need the prime framework...you need orchestration!

00:18:08: ...You can't just let twenty agents run wild without a central nervous system like Salesforce to hold all of those contexts together.

00:18:15: And the visibility piece?

00:18:16: It's not about Google anymore…It's ensuring.

00:18:19: when a developer asks for coding bot or CEO asking perplexity your brand is the recommended answer—not passing mention.

00:18:29: What stands out to me the most across all these sources is this recurring theme of infrastructure before action.

00:18:34: Whether it's Spencer Perique talking about cleaning your data, before emailing or Matt Hines talking about orchestration Or Annie York fixing The About Page thema?

00:18:42: The winners in twenty-twenty six aren't the ones with the flashiest new AI tool.

00:18:46: They're the one who have done the boring unsexy work Of building a system that allows AI To function correctly.

00:18:51: That Is THE BIG AHA MOMENT FOR ME!

00:18:54: AI Just amplifies what you already are.

00:18:56: If your processes are messy, you just become messy at a massive scale.

00:19:00: if You're structured...you become an absolute machine.

00:19:02: Precisely And for you listening I think the big takeaway is to audit Your own role.

00:19:07: today Are you doing work that A specialized agent could do For twenty-five dollars a month?

00:19:14: Or are you Doing The Work of the human moat The relationship building The strategic orchestration The creative founder led storytelling?

00:19:23: Because that line between what the machine does and you do is moving every single week.

00:19:29: It is a little scary, I won't lie but it's mostly exciting.

00:19:57: marketing, martech go to market and social selling.

00:20:01: Thanks for listening.

00:20:02: everyone don't forget to subscribe.