Best of LinkedIn: AI in B2B Marketing CW 11/ 12

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 in calendar weeks eleven and twelve.

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

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

00:00:21: Welcome

00:00:22: To The Deep Dive.

00:00:24: So what if I told you tomorrow?

00:00:27: The traditional SEO playbook you've spent the last ten years perfecting.

00:00:31: And probably hundreds of thousands of dollars executing, let's be honest!

00:00:34: Right exactly what if I told you that entire Playbook is just going to completely useless because today we are basically tearing down the way BtoB marketing has operated for the past

00:00:43: decade?

00:00:43: Yes complete tear-down.

00:00:44: We're looking strictly at most insightful truly disruptive trends curated from LinkedIn during calendar weeks eleven and twelve... ...and were translating all that noise into pure signal for your strategy.

00:00:54: Exactly we are cutting past the generic hype to focus on three massive structural shifts happening in the market right now.

00:01:05: We're going to examine the sudden evolution of AI search visibility, then we'll break down mechanics of signal-based outbound and finally will explore the reality of a gentic go-to-market strategies an autonomous workflow

00:01:16: design.".

00:01:16: Okay let's untack this because before you can even think about automating your sales outreach or designing a brilliant marketing campaign... Your buyers actually have to find you!

00:01:27: Right

00:01:27: top.

00:01:28: Yeah, that top of funnel discovery phase is undergoing a seismic shift right now.

00:01:33: We are moving away from traditional SEO or search engine optimization towards something fundamentally different.

00:01:38: we're talking AEO Answer Engine Optimization Or GEO Generative Engine Optimisation.

00:01:44: What's fascinating here Is the underlying mechanism Of how a buyer discovers B to be solution Has just fundamentally changed.

00:01:51: Samantha Stewart recently shared a really brilliant analysis on this shift.

00:01:54: She pointed out that AI It simply does not show you ten blue links to choose from.

00:01:59: No, it doesn't give you options to click through

00:02:02: right?

00:02:02: It synthesizes the information and provides a direct singular recommendation.

00:02:07: So the legacy question every marketing team used ask was how do we rank on page one of Google?

00:02:12: but The new much harder question is will an AI model actually recommend my brand over my competitor?

00:02:18: Okay I have to play devil's advocate here.

00:02:20: isn't this just SEO with a trendy new acronym?

00:02:22: i

00:02:22: mean A lot people think that yeah

00:02:24: Because if the goal is just getting an AI to read your website instead of Google's crawler, marketers are going pump out hundreds of AI-generated blog posts.

00:02:34: Stuff them with the right keywords and try to trick chat GPT, right?

00:02:38: It feels like trying a cram for open book test but bringing entirely wrong

00:02:45: textbook.".

00:02:47: George Cushteria recently shared some hard numbers on where large language models actually pull their answers from.

00:02:52: So this was wild?

00:02:54: It really is, when a BDB buyer asks perplexity or chat GPT for say the best CRM software The number one sources those model site are Reddit at eleven point two nine percent and LinkedIn at eleven Point zero three percent.

00:03:08: wait reddit above actual corporate websites

00:03:12: Above corporate websites above major news syndicates mm-hmm about Wikipedia.

00:03:17: Those human-to-human forum platforms are sitting way above your perfectly optimized corporate landing page.

00:03:24: Okay, but I want to challenge that.

00:03:25: yeah because if Reddit and LinkedIn or the most trusted sources with the highest algorithmic waiting Won't BtoB marketers just pivot their automated bot farms to spam Reddit threads?

00:03:36: That's the instinct, sure.

00:03:37: Like if I know that AI is reading a subreddit about enterprise software... ...I could deploy ten bots and have a fake conversation on how great my product was!

00:03:44: Why wouldn't it

00:03:45: work?!

00:03:45: Because that assumes AI visibility as a content output

00:03:48: problem

00:03:49: which

00:03:50: hmmm....

00:03:51: It isn't.

00:03:52: Julian and Cigarys Miller specifically addressed this, she argued that AI visibility is actually a trust distribution problem.

00:03:58: A trust distribution problems?

00:03:59: Okay explain them.

00:04:00: Think of your brand's reputation like a financial credit score.

00:04:03: it no longer matters how loudly you shout your own excellent credit score on your own website.

00:04:08: The AI model is essentially checking with fifty independent banks across the internet to see if your checks are clear.

00:04:15: Oh!

00:04:15: That's great way to

00:04:16: put it Right.

00:04:16: So If You Spam Reddit With Fake bought driven reviews.

00:04:21: The models recognize the lack of genuine historical corroboration, they know it's

00:04:26: synthetic.".

00:04:27: And Miller cited that startup tally as a perfect case study for this right?

00:04:31: Yes!

00:04:32: They are an eight-person form building company—eight people and they're consistently beating massive heavily funded legacy incumbents in AI recommendations.

00:04:43: And how did they do it?

00:04:44: Their founder, Marie Martin spent literally years organically building authentic public trust in indie developer communities and on Reddit.

00:04:52: She wasn't just dropping promotional links right.

00:04:54: she was genuinely solving problems.

00:04:56: Exactly!

00:04:57: She's answering questions.

00:04:58: being a real human.

00:04:59: Miller calls that crossing the corroboration threshold.

00:05:03: It's the moment when enough independent credible sources, real humans with historical account authority say the same positive things about your brand.

00:05:11: And once you cross that threshold The AI commits to recommending

00:05:15: it And you absolutely cannot fake that with an automated blog farm or a sudden burst of synthetic Reddit comments.

00:05:21: That requires the total philosophical shift for marketing teams, I mean Bill Habeb advised that teams have to move away from traditional content marketing and start building content systems.

00:05:31: Content systems exactly!

00:05:32: Because standard rambling two thousand word blog post might've worked for Google keyword density but it is terrible for an LLM trying to extract quick fact.

00:05:43: You have structure your data so AI can actually parse it, real case studies verifiable numbers clear problem and solution architectures.

00:05:51: And Damon Pestulka captured the essence of this transition perfectly.

00:05:55: he basically said clarity wins period.

00:06:00: AI models natively boost businesses that can consistently and simply explain customer problems, and their subsequent solutions.

00:06:07: The models ignore the company's hoarding complex tech jargon in an attempt to sound you know sophisticated

00:06:12: Right.

00:06:13: trying to sound smart actually hurts You.

00:06:14: now exactly

00:06:15: To be the recommended choice?

00:06:16: Yeah have to be the clearest voice in the room not the loudest.

00:06:19: That is so true And I think that sets up our second major shift perfectly.

00:06:23: Let's assume you've built that trusted visibility you've crossed the corroboration threshold, and buyers are starting to enter your orbit.

00:06:30: The next logical step for a B-to-B team is reaching out.

00:06:33: The Outbound Motion

00:06:34: Exactly!

00:06:35: Outbound.

00:06:36: But looking at insights from calendar weeks — eleven and twelve—the current instinct is completely backwards.

00:06:43: Teams use new AI tools to blast emails.

00:06:46: literally everyone with a pulse…and the smartest organizations run in exactly opposite directions.

00:06:52: If we connect this to the bigger picture, We're witnessing what Yuri Zaremba accurately described as The tragedy of the AISDR market.

00:06:59: It really is a tragedy!

00:07:00: The vendor landscape right now Is locked in a race To the bottom.

00:07:03: Everyone is promising to send More outbound messages Faster and cheaper And the mathematical result Of that race Is flooded executive inboxes And completely collapsed response rates Across the entire B-to-B sector.

00:07:16: Yeah Kevin N shared a hilarious Well, hilarious but incredibly depressing observation about this.

00:07:22: He looked at the new AISDR patents being filed right now and The underlying technology is literally just generate email send email auto reply.

00:07:31: that's it That there was no strategic layer freedom to allow put appointment

00:07:36: taking the spray-and-pray approach that already wasn't working in twenty, twenty three and giving it a high pressure fire hose.

00:07:43: Yeah you're not innovating.

00:07:45: You were just automating mediocrity at a terrifying scale right?

00:07:48: And when you scale spam The communication channel dies.

00:07:51: but the viable alternative emerging from the smartest operators is signal based outbound.

00:07:57: Nyla Gorman pointed out A fundamental truth about machine learning AI will relentlessly optimize whatever data.

00:08:03: Okay, so if you feed it garbage

00:08:04: and get optimized garbage.

00:08:05: If you feed the system basic activity metrics like email opens or link clicks It will optimize.

00:08:10: to get your more meaningless activity You have to feed the AI actual change signals.

00:08:14: Let's clearly define the mechanism of a changed signal versus an Activity metric.

00:08:18: just so we're all on the same page.

00:08:20: An activity metric is like someone accidentally leaving an e-mail open On their second monitor while they go to lunch

00:08:25: right?

00:08:25: He's nothing

00:08:26: but a chain signal.

00:08:27: Is a new vice president of operations joining your target account or a company publicly announcing a new round of funding, Or a sudden concentrated spike in hiring for specific technical role.

00:08:39: Those are moments that indicate structural intent.

00:08:43: And Gouraj Panarangi summarized this as the shift from scaling effort to scaling intelligence.

00:08:49: instead of blasting a purchase list of ten thousand names and hoping for one percent conversion you're executing intent-driven outreach.

00:08:58: The AI is monitoring the web for those specific change signals and drafting a highly contextual message, the exact moment that signal occurs.

00:09:06: So what does this all mean?

00:09:08: For RevOps leader if you're listening to this And your under pressure to cut costs This quarter Your immediate instinct might be To just buy an off-the-shelf AISDR tool today Point at your CRM fire half your junior sales team and let the software rip.

00:09:25: Please don't do that.

00:09:26: But why shouldn't a company just do that?

00:09:28: It sounds

00:09:28: cheaper.

00:09:29: Because the failure rate of that exact strategy is catastrophic.

00:09:33: Edward Gorbis highlighted recent Deloitte study, which serves as massive reality check for Revox leaders.

00:09:39: Between fifty and seventy percent of AISDR tools churn within first twelve months.

00:09:45: Fifty to seventy

00:09:47: percent churn.

00:09:48: Yes, and the underlying software isn't necessarily broken.

00:09:51: The CRM data it relies on is a complete disaster.

00:09:54: Oh

00:09:54: right because over sixty eight percent of critical sales data lives outside the CRM in random spreadsheets lack messages or you know sticky notes On someone's monitor

00:10:02: exactly.

00:10:03: amine slam made a profound point regarding this dynamic AIS.

00:10:06: DRs are essentially just cloning machines.

00:10:08: They do not invent strategy.

00:10:10: they clone what has already there.

00:10:11: So if your CRM is filled with duplicate contacts, people who left their jobs three years ago and a totally undefined ideal customer profile.

00:10:21: The AI will perfectly clone you're broken human process.

00:10:25: it will execute Your bad strategy flawlessly feeling much faster And burning your domain reputation in the process.

00:10:32: Jonathan MK refers to this as the mandatory foundation fix

00:10:36: Foundation Fix.

00:10:36: I like that.

00:10:37: AI will actively destroy your pipeline strategy if you deploy it on top of unmapped processes and messy data.

00:10:44: The organizations actually seeing ROI from AISDRs are the ones that did incredibly boring, unglamorous work.

00:10:51: first

00:10:52: They clean their rooms.

00:10:53: Yes they de-duplicated their records, clearly defined their ICP And mapped to exact workflows before ever introducing an automated agent.

00:11:01: And that foundation fix is the absolute prerequisite for the third and most significant shift we found in research.

00:11:07: Agented go-to market, an autonomous workflow design because once your data is pristine and you're operating on signal based intent AI stops functioning as a simple autocomplete tool.

00:11:17: it transitions into an autonomous teammate.

00:11:20: Charlie Saunders provided great framework to understand this evolution.

00:11:24: He talks about moving from level two AI assistance to Level three and level four production grade AI automation.

00:11:31: Right, let's break that down for the listener

00:11:33: To understand the mechanics of this picture a level-two workflow.

00:11:37: You manually copy a prospects LinkedIn profile you paste it into Claude?

00:11:41: You prompt the AI to write a personalized cold email.

00:11:44: you review The text and then you manually paste it back in your outreach tool.

00:11:48: So the AI assisted you, but are still operational bottlenecks?

00:11:52: Exactly.

00:11:53: You're doing heavy lifting!

00:11:54: Right whereas a level four system operates completely autonomously in the background The agent monitors LinkedIn for job change recognizes signal queries your data warehouse with that specific prospect drafts highly contextual email and sends it all while literally asleep.

00:12:13: And underlying economics driving this transition is undeniable.

00:12:17: Matthew B. Borner broke down this stark mathematical reality, a fully loaded human sales development representative factoring in base salary commissions benefits software licenses and management overhead costs the company between one hundred thousand and one hundred sixty five thousand dollars a year.

00:12:34: okay and an AI

00:12:36: dedicated AI sales agent running continuously cost between six thousand twenty four thousand dollars.

00:12:41: That is an eighty five to ninety-five percent cost reduction for top of funnel execution.

00:12:46: It's insane!

00:12:48: And it isn't just theoretical, Jason M Lemkin the founder of Sackstruck recently discussed how a company essentially replaced ten person go-to market team with one point two human operators managing twenty AI agents.

00:13:00: One Point Two Humans.

00:13:01: Yeah and they completely maintained in some metrics actually improved their business performance.

00:13:05: This scalability extends well beyond enterprise tech companies.

00:13:09: David Sold documented how he built a fifteen-roll AI operating system to run his entire agency.

00:13:14: He deployed specialized agents acting as an AI chief marketing officer, and AI Chief Operating Officer ,and an AI Chief Revenue Officer.

00:13:22: Yeah!

00:13:22: And Gow's Masawi highlighted in open source framework called Open Claw that takes this to the extreme – it is an AI agent running on basic seven dollar month server.

00:13:30: Seven dollars?

00:13:31: I mean if you are listening right now looking at your six figure marketing tech stack….

00:13:36: It really absorbed that, a seven dollar server.

00:14:00: But notice the built-in restraint of.

00:14:04: it is programmed to only send two emails per prospect.

00:14:07: It doesn't harass them with a seven-step sequence because it utilizes historical data proving that after two unanswered e-mails, you're simply trashing your domain

00:14:15: deliverability.".

00:14:16: Right!

00:14:16: That's an autonomous agent making a strategic, data driven decision not just executing volume play.

00:14:21: Here's where gets really interesting though Because there was massive systemic flaw in how most companies are currently setting up these Autonomous Agents.

00:14:30: Jerry Farr identified as AI Amnesia.

00:14:33: Yes

00:14:34: AI amnesia.

00:14:35: If you don't build the architecture correctly, your brilliant agent suffers from a severe memory deficit

00:14:41: far compared to current state of most tools.

00:14:44: to the movie Groundhog Day.

00:14:45: Oh

00:14:46: I love this analogy.

00:14:47: right now When a marketer opens the new session in an AI tool, it starts completely from scratch.

00:14:53: It forgets that deep account narrative distilled yesterday and gets the nuanced tone of voice that finally nailed last week.

00:15:00: The AI wakes up every single day with zero context And

00:15:04: if you remember Groundhog Day... ...the only reason Bill Murray's character eventually escapes time loop is because he kept his memories.

00:15:10: He learned to play piano by compounding knowledge over thousands days.

00:15:14: If your AI agents are constantly waking up on February second with zero historical context, they can never improve.

00:15:20: Exactly and this requires a structural solution which far calls the shared memory hub.

00:15:26: He advises placing a central data warehouse like Snowflake or BigQuery at the absolute center of your go-to market stack.

00:15:33: Before an AI agent executes a task, it pings the data warehouse to pull historical context

00:15:38: And then after the agent finishes writing and email or analyzing a call It writes its findings back into that Central Hub.

00:15:45: Exactly!

00:15:46: The mechanism there is compounding intelligence.

00:15:49: If your AISDR is drafting a Q-three outreach email, it first checks the memory hub and realizes that an AI research agent flagged a relevant Q one earnings call transcript for that exact account months ago.

00:16:01: So the q three email becomes hyper personalized based on Q One data.

00:16:05: right if you're AI system has compiling memory in Your competitors AI as daily amnesia You are going to win the account every single time absolutely.

00:16:13: But this raises an important question regarding the human element.

00:16:15: yeah As we transition into an era of agents swarms dozens of specialized memory-enabled agents handling research, signals outreach and analytics.

00:16:24: What is the actual role of a human marketer or seller?

00:16:27: Yeah that's existential anxiety vibrating through the B to B sector right now.

00:16:32: Are we all just destined for becoming software babysitters?

00:16:36: Mark Serkin provided brilliant framing.

00:16:39: He recently received a highly personalized cold email generated by an AI.

00:16:45: The agent successfully scraped his name, location and recent activity….

00:16:50: the grammar was flawless.

00:16:51: Okay

00:16:52: sounds good so far!

00:16:53: However, he completely hallucinated his business category making the core pitch entirely irrelevant.

00:17:00: Serkin noted that AI can perfectly replace execution but it utterly fails at replacing judgment.

00:17:06: He argued that AI makes operators lazy until they fail publicly.

00:17:09: The AI basically borrows its underlying hypothesis from the mathematical average of everything.

00:17:35: From an execution standpoint, it was incredibly fast and cost efficient.

00:17:39: But from a market standpoint

00:17:40: the audience completely rejected It.

00:17:42: they described the campaign as soulless Noting it had lost all the historical brand warmth that defined Coca-Cola.

00:17:49: Wow And you have to contrast that failure with unilever.

00:17:53: Unilever is succeeding with AI at scale because they refuse to let the models run wild on execution.

00:17:59: They built a heavy, mandatory human strategic layer on top of technology.

00:18:03: Yes!

00:18:04: The human layers are critical.

00:18:05: They created an internal system called brand DNAI.

00:18:08: Every single AI-generated asset must pass through this system which is strictly managed by humans acting as prompt strategists and AI creative directors.

00:18:16: The humans ensure that output aligns with approved brand voice cultural context emotional resonance

00:18:22: Because the humans are no longer executing the tedious busy work of formatting spreadsheets or drafting introductory paragraphs.

00:18:29: They're injecting the empathy, cultural awareness and creative direction.

00:18:33: because AI possesses zero emotional intelligence

00:18:37: exactly.

00:18:38: it can optimize a process and predict probability but cannot feel.

00:18:42: The marketers who will dominate next decade won't be ones that know how to operate most software tools.

00:18:48: they'll leaders understand human judgment to machine scale.

00:18:53: So what does this all mean?

00:18:55: Which brings us to a final, provocative thought for you to mull over as you design your strategy for the rest of the year.

00:19:01: Elizabeth Hughes shared an insight that fundamentally reframes this entire technological shift.

00:19:07: right now every B-to-B company is rushing to sell their newly minted AI powered solution.

00:19:12: but the biggest go-to market shift of this decade isn't actually about sales funnels or feature lists.

00:19:17: it's about education.

00:19:19: Education overselling

00:19:19: Exactly, your buyers are overwhelmed.

00:19:23: AI implementations don't fail because the software is too expensive?

00:19:26: They fail because buyers literally do not understand how to integrate these tools into their daily workflows without breaking their businesses!

00:19:35: If

00:19:35: A.I.,

00:19:36: is a new operating system for the commercial world teaching your buyer's how-to survive and thrive in it is your single most powerful distribution channel.

00:19:43: Are you just aggressively selling software to market or actively educating them?

00:19:48: That's a powerful question to end

00:20:07: on.