Best of LinkedIn: MarTech Insights CW 26/ 27

00:00:00: This episode is provided by Thomas Allgaier and Frennis, based on the most relevant LinkedIn posts about MarTech from calendar weeks twenty-six and twenty seven.

00:00:08: Frenis is a B to B market research company that supports enterprise marketing teams in unlocking the full potential of their customer data with the help of AI.

00:00:16: you can find more info.

00:00:19: So what if the millions of dollars your company just spent painstakingly integrating its marketing technology stack is you know exactly What's going to make?

00:00:27: Your entire department structurally obsolete in two years?

00:00:29: man that Is the multi-million dollar question keeping chief marketing officers awake at night right now.

00:00:36: yeah, and it's Exactly.

00:00:37: We were exploring today.

00:00:38: Welcome back to the deep dive, everyone.

00:00:39: We're looking at a fascinating stack of curated insights from top BDB marketing professionals over on LinkedIn.

00:00:46: and really the overarching theme here is this radical shift from buying isolated software tools.

00:00:55: Right, and our core mission today is to essentially distill all that noise because there's a lot of hype around agentic AI vendor consolidation data architecture.

00:01:05: We want to figure out what's real?

00:01:06: What's just vendor spin And how you can actually prepare for reality where the rules of software have just fundamentally changed.

00:01:15: Yeah Let's jump right into The most aggressive claim we found which Is that the very concept Of a martech stack is Just

00:01:22: dead.

00:01:22: Yep completely dead.

00:01:24: Rupert Steffner brought this up.

00:01:26: He argues that competitive advantage no longer comes from the tools you happen to own but form transforming marketing into what he calls a value discovery and activation flywheel, basically AI operating model.

00:01:39: Right,

00:01:39: and the distinction between a stack in a flywheel is I mean it's entirely about mechanics.

00:01:43: think about a traditional software stack like A Tower of Blocks right?

00:01:46: Yeah You buy a marketing automation platform you stick a webinar tool on top of that maybe slide CRM underneath but It just sits there waiting for a human to pull the levers

00:01:55: right.

00:01:55: its passive

00:01:56: exactly.

00:01:57: But a fly wheel when it's driven by AI agents Is structurally different because it's autonomous.

00:02:03: It continuously mines your data, finds a buying signal spins up the campaign to capture it measures.

00:02:10: And then and this is key feeds that learning back into the system so its spends faster than next time.

00:02:15: That

00:02:16: sounds amazing but you can't just like go shopping for flywheel right?

00:02:20: No definitely not.

00:02:21: Jane van doesburg made a brilliant point about This.

00:02:23: she said You have to design The business operating model before Look at a single piece of technology.

00:02:29: Oh, absolutely

00:02:30: like ask yourself where the business wants to be in five years and figure out what decisions should Be automated versus What actually needs a human.

00:02:37: because if you don't do that?

00:02:38: If you don t define The Business Logic first any AI you buy is just going To automate your current dysfunction

00:02:44: doing the wrong things lightning fast

00:02:46: exactly.

00:02:47: And this ties perfectly into what John Miller calls reasoning ai.

00:02:51: he pointed Out That legacy marketing automation platforms are basically Just rigid.

00:02:56: I have thin rules engines.

00:02:58: Oh yeah, MOPs teams it's been their whole lives writing that.

00:03:02: uh we called spaghetti logic.

00:03:05: yep spaghetti like if a lead fills out this form at a point.

00:03:08: they live in California route them here.

00:03:11: And the fatal flaw there is that human data is just incredibly messy.

00:03:15: A buyer might type CA or PAL, or caliph into a form and in a legacy system if you didn't manually write a rule for every single spelling The system breaks!

00:03:26: The lead is lost.

00:03:27: Reasoning AI changes it entirely because uses large language models to interpret ambiguity

00:03:33: So actually understands context.

00:03:35: It looks at a profile and intrinsically knows that A Director of Operations At a Logistics Company is totally different from A Director Of Operations.

00:03:42: At a Software company, it doesn't need a human to write specific routing rule for every variation.

00:03:47: Okay

00:03:48: I hear That but i have To push back a little here.

00:03:50: Are we just...I mean isn't an Operating Model Just a fancy Consultant-y way of saying a well organized tech stack?

00:03:56: are We just rebranding the same old software to sound futuristic ?

00:04:00: I get why you'd say that it definitely sounds like a rebrand on The surface.

00:04:04: But Ahmed da two mapped out this six-layer marketing operating system to show why it's a structural evolution, not just a label.

00:04:13: Okay walk me through that.

00:04:14: So these layers are specifically designed to give AI agents the safe environment they need.

00:04:18: at The bottom you have the data and identity layer.

00:04:20: That's where the system figures out who is this person?

00:04:23: Right merging all those behavioral signals

00:04:25: exactly?

00:04:27: Then layer two is life cycle which tells the AI where they are in the buying journey.

00:04:31: Layer three is policy And this one is massive.

00:04:35: Well, because that's the guardrails right?

00:04:37: Yes

00:04:38: policy houses your ideal customer profile Your compliance rules.

00:04:41: it literally tells the AI you can offer a discount here But absolutely not to this other group

00:04:46: okay.

00:04:47: So that keeps the AI from going rogue.

00:04:49: precisely then above That is Decisioning The brain that picks next best action.

00:04:54: Then omni-channel Action which actually sends the email and finally learning in measurements sitting on top.

00:04:59: Wow, okay.

00:05:00: So it's this perfectly sequenced pristine six-layer sandbox.

00:05:04: but what happens when you plug those agents into the duct?

00:05:07: taped messy tools most companies actually have right now?

00:05:10: Oh!

00:05:10: It's just complete chaos.

00:05:12: Karsten Lackner issued a stark warning about this.

00:05:14: he said adding AI to a messy stack doesn't fix it.

00:05:18: AI inherits your data...it doesn't magically solve

00:05:21: Right, so you're just empowering your company to be wrong at scale.

00:05:24: Exactly!

00:05:24: If your CRM thinks a customer is active but your billing software knows they churned the AI will send highly personalized retention emails to someone who's actively suing you.

00:05:36: Oh man that was a nightmare.

00:05:38: and it really brings us into what Prof Alex Varsay calls integration tax or data latency penalty.

00:05:44: Yes

00:05:44: this is huge mechanical problem.

00:05:46: Because for years the gold standard was building a best of breed stack by the Best Webinar tool, The Best Email Tool.

00:05:53: and just sketch them together with custom APIs.

00:05:55: but those API bridges cause massive data delays.

00:05:58: they are not instantaneous.

00:06:00: right.

00:06:00: you have to extract the data translate it load it yep.

00:06:03: so imagine if prospect hits your pricing page Right now showing huge intent By the time that signal goes from the web tracker across the API into CRM, waits in a queue and syncs to outreach tool.

00:06:17: Fifteen minutes have passed!

00:06:18: Exactly And In an AI world.

00:06:21: fifteen minutes is eternity.

00:06:23: Your best-of-breed advantage is gone because a competitor with natively integrated platform already engaged them.

00:06:29: Which is why Jeanine Ferguson updated her MarTech framework to make data standardization mandatory prerequisite for AI?

00:06:36: You can't just skip that step anymore!

00:06:37: No, you really cannot and I love Pierre Ligori's metaphor of this – he compared MarTech platforms to pasta…and marketing data into the sauce.

00:06:45: Oh yeah...this was great.

00:06:46: People are obsessing over buying those expensive real time pasta machines

00:06:50: But they're powering it with packet sauce.

00:06:52: Most companies overcook the pasta and use packet sauce,

00:06:55: right?

00:06:55: If your data relies on manual spreadsheet uploads you just bought a multi-million dollar way to serve bad pasta.

00:07:01: Yeah To be confidently wrong.

00:07:03: It totally reminds me of Phil Gamache an anim arounds podcast series where they turned Martek architecture into this RPG style dungeon

00:07:12: crawl.

00:07:12: Oh I saw that!

00:07:13: Its'a brilliant way to look at governance.

00:07:15: yeah

00:07:15: They had these bosses.

00:07:16: You have to fight like The Hallucination Oracle or Rotten context mage, which is basically just orphaned data from a twenty nineteen webinar.

00:07:25: Exactly and Anna Moran pointed out that the only way to defeat those bosses Is with the data template potion.

00:07:31: but it can't Just be a static word document on a shared drive

00:07:34: right?

00:07:34: It has to Be living governance.

00:07:36: yeah enforce at The code level because without it your AI just produces as she put it believable nonsense faster

00:07:43: Believable nonsense.

00:07:44: Yeah That is so dangerous Because executives love things that sound confident.

00:07:48: So If data latency and these API bridges are killing AI performance, it totally makes sense that vendors are racing to own the decisioning middle layer.

00:07:57: Oh they were aggressively trying to eliminate those bridges.

00:07:59: Jonathan Moran pointed this out The race isn't about email delivery anymore.

00:08:03: Every vendor wants be engine decides what to send when.

00:08:07: Yeah look at Databricks launching customer lake.

00:08:09: That was a blockbuster move.

00:08:11: Right because their back end data warehouse company.

00:08:14: But Customer Lake is an agentic CDP built natively into the warehouse

00:08:19: And Yolanda DeMello noted how this just completely blurs the line between a data warehouse and MarTech.

00:08:25: If the CDP lives inside the warehouse, there's no API bridge—no latency

00:08:30: penalty.".

00:08:31: So you can run these infinity campaigns at millisecond speeds?

00:08:34: Yes!

00:08:35: But secret DRX offered massive reality check here —the market narrative assumes companies actually have governed data warehouses in perfect identity resolution...

00:08:45: Right like everyone is ready to go… Exactly.

00:08:47: But the market reality is most mid-market and enterprise companies are still trying to get their CRM, an eCommerce platforms to agree on a single email address.

00:08:57: They're still trying build there very first consistent customer view.

00:09:01: but the vendors don't care.

00:09:02: they were forcing this anyway.

00:09:03: Christopher Marriott was talking about his huge M&A frenzy.

00:09:07: That's

00:09:07: wild right now!

00:09:08: Moe engage buying ampi sales force buy in Contentful Blue Connick grabbing blue shift.

00:09:13: Yeah, everyone is buying up whatever pieces they lack to build a single orchestration layer.

00:09:18: But wait... if Salesforce's buying up AI agents and headless CMS platforms aren't we just going right back the giant monolithic Franken-sweets that we tried to escape ten years ago?

00:09:30: It's

00:09:30: very valid concern.

00:09:31: Florian Delvall warned about this exactly saying that The Bundle Is Back in history basically repeating itself.

00:09:38: So were trapped in legacy bundles again.

00:09:40: Well, their goal this time is to become a single engagement operating system.

00:09:44: So instead of just duct taping acquisitions together with internal APIs they want to force everything onto one underlying data model.

00:09:51: Okay so that actually solves the integration tax?

00:09:53: It does but the trade-off is vendor lock in.

00:09:56: If you put all six layers on one vendors unified model You are never leaving

00:10:01: Right!

00:10:01: Your'e trading the integration for the lockin tax.

00:10:05: With all these technology consolidating and AI taking over the execution layer What does an actual B-B marketing professional do now?

00:10:13: It's a

00:10:13: massive shift.

00:10:14: The focus moves away from plumbing and goes straight to the people in the context they provide.

00:10:20: Christopher Swarup argued that, MarketingOps has to evolve from workflow builders into a GTM control plane

00:10:27: Before agents are even allowed touch anything.

00:10:28: Exactly!

00:10:29: And Daryl Alfonso actually mapped out four future mom ops roles for this.

00:10:34: First is AI engineer

00:10:35: Okay...the builder

00:10:36: Right.

00:10:37: Then the RevOps data lead to build the context and governance layer, then uh...the Agentops manager

00:10:44: Which is such a wild job title It

00:10:45: IS!

00:10:46: But they monitor the fleet of agents adjust guardrails make sure that aren't hallucinating And finally The RevOps analysts run this system and measure outcomes

00:10:56: Getting humans out of manual execution business honestly for their best.

00:11:00: Sarah McNamara shared this war story where she audited an enterprise system and found eleven different spellings of the word campaign.

00:11:07: Oh wow, just human error at scale?

00:11:09: Exactly!

00:11:09: So she advocates for centralized production systems like NACC.

00:11:13: so MOP stops being the cleanup crew constantly reconciling channels after launch.

00:11:18: Right if AI handles a deployment from a rigid brief The Human is freed up.

00:11:23: And Lea Cassidy and Janessa Lance highlighted exactly what they should be doing with that time building messaging architecture.

00:11:30: Okay, wait what does that actually mean?

00:11:32: Because it can't just be a PDF about brand tone.

00:11:35: No, AI can't read a vibe from a pdf.

00:11:37: It needs structured machine readable context.

00:11:41: marketers have to build these dense matrices of products differentiators specific use cases.

00:11:46: Oh I see.

00:11:47: because if you don't feed at that structured architecture the AI Just spits out generic robotic garbage.

00:11:53: exactly That architecture is your new competitive edge?

00:11:56: It sounds like marketers are basically moving from being The line cooks frantically chopping vegetables in the back to being the executive chefs.

00:12:04: That is a perfect analogy!

00:12:06: Like

00:12:06: you design the menu, source ingredients via context while AI does actual repetitive cooking

00:12:12: And it aligns beautifully with what Cecily B took away from the ConLions Festival.

00:12:15: Despite all this talk about agentic AI and composable MarTech The ultimate goal was making ads for humans Not algorithms or spreadsheets.

00:12:24: Right because buyers are still stressed out Humans trying solve problems.

00:12:28: Exactly Rene as noted that when founders pitch the one who wins.

00:12:32: The room isn't the one with the best tech stack It's the one, Who is the most authentic?

00:12:37: Authenticity beats tech every single time because algorithms fundamentally lack empathy.

00:12:42: They

00:12:42: don't have lived experiences.

00:12:44: right

00:12:44: and finding That human layer understanding the emotional toll of the buyers problem That's a data set technology cannot generate.

00:12:52: So, as AI makes execution instantaneous and flawless across the board.

00:12:56: And every competitor has the exact same AI capabilities.

00:13:00: Efficiency just becomes a complete commodity?

00:13:03: Yeah

00:13:03: so the only true competitive moat left for B to be brand is going to that unautomatable messy distinctly human intuition.

00:13:12: in creativity you feed into machine first

00:13:14: place Absolutely!

00:13:15: The machine can cook but have give it taste.

00:13:18: If you enjoyed this episode, new episodes drop every two weeks.

00:13:21: Also check out our other editions on account-based marketing, field marketing channel marketing AI and AAB.

00:13:27: to be marketing go to market in social selling.

00:13:30: Thank you so much for joining us On This Deep Drive today.

00:13:32: Don't forget to subscribe And we will see all next time.