Best of LinkedIn: MarTech Insights CW 12/ 13

00:00:00: This episode is provided by Thomas Allgaier and Freeness based on the most relevant LinkedIn posts about MarTech from calendar weeks, twelve and thirteen.

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

00:00:17: you can find more info in the description.

00:00:20: it's really great to be here.

00:00:21: yeah we have a massive deep dive today.

00:00:23: I mean imagine lives for exactly five minutes to complete one, you know highly specific operational task and then it just completely deletes itself.

00:00:35: Yeah

00:00:35: no saw subscription No integration headache

00:00:39: right zero residual tech debt.

00:00:42: So today we're diving into the absolute top martech trends that We've been seeing buzzing across LinkedIn for B to be marketing professionals.

00:00:50: and look, we are moving way past The fluffy buzzwords Today.

00:00:53: Oh definitely

00:00:54: were getting Into the structural shifts

00:00:55: exact.

00:00:56: I mean if you're a marketing director listening To this on your commute right now You're probably sweating over there in A half-dozen sauce.

00:01:02: renewals?

00:01:02: You have coming up next quarter

00:01:04: And you're Probably wondering If you should Be renewing them at all.

00:01:07: Right

00:01:07: because the fundamental architecture of Martek is collapsing.

00:01:11: AI agents are stepping in to take over not just the marketing side, but really the entire B-to-B buying

00:01:17: process.

00:01:18: Yeah and the underlying theme across all these insights we analyzed from the past two weeks.

00:01:22: is that your competitive advantage?

00:01:24: well It's no longer gonna come from just buying another commercial off-the-shelf software tool

00:01:29: Which is how?

00:01:29: We've done it for ten years exactly.

00:01:31: The new game is entirely about data architecture coordination.

00:01:38: Just relentless execution discipline.

00:01:41: Wait, okay.

00:01:41: So are you telling me that these enterprise companies who have spent literally millions of dollars over the last decade?

00:01:47: Bolting Marketo to Salesforce a snowflake?

00:01:50: They're just gonna throw that architecture in the trash?

00:01:52: well no they won't throw it out overnight But they are absolutely shifting where the center of gravity is.

00:01:57: I mean Scott Brinker just released this massive new report Co-created with Databricks declaring what he calls The Third Age Of Martech

00:02:04: the third age.

00:02:05: Okay

00:02:06: Yeah, the traditional model where you have these multiple rigid layers of sauce tools just stacked on top of each other that's dying.

00:02:14: It's being replaced by what he calls a composable canvas and this canvas is built directly on top Of a unified data foundation like uh Like a data lake house.

00:02:24: hold on before we go further.

00:02:25: for those of us listening who aren't You know data engineers?

00:02:28: What actually is a lakehouse compared to the standard CRM database We've been forcing our teams to use for a decade.

00:02:36: Right, okay.

00:02:36: So think of your traditional CRM database as a highly organized filing cabinet.

00:02:41: Okay It's great for specific structured data But to do anything complex you basically have to photocopy those files and mail them To thirty different departments.

00:02:49: And

00:02:49: those apartments are all your different sauce tools.

00:02:51: Exactly it's messy.

00:02:52: A lake house is more like a massive flat reservoir That holds All your company's Data in its raw form.

00:02:58: Oh I see

00:02:59: Yeah.

00:02:59: so instead Of moving data across disjointed tools which you know causes immense latency and integration nightmares the data just stays put.

00:03:07: Okay let's unpack this.

00:03:09: so instead of trying to bolt together thirty different prefab sheds in your backyard, hoping they all share plumbing we're basically pouring one massive concrete foundation and letting AI instantly build, and tear down whatever rooms we need exactly when we need them.

00:03:24: That is a perfect analogy.

00:03:26: the applications The workflows that AI agents they all just sit directly on top of that shared reservoir And sip the data They need in real time

00:03:34: okay?

00:03:35: That makes sense conceptually but what does the actual business impact of that?

00:03:38: I know.

00:03:38: Alexander Burton and Felix Santiago were both highlighting this shift on LinkedIn recently.

00:03:44: They pointed out that Databricks' own marketing team actually moved their lead scoring model in-house onto this exact architecture.

00:03:50: And what happened?

00:03:51: The result was a forex jump in their conversion rates compared to the external vendor they replaced.

00:03:56: Wait,

00:03:56: I have to push back on that!

00:03:58: Why does simply moving a lead scoring models in house into lake house quadruple

00:04:04: conversions?!

00:04:05: Is third party software really bad?

00:04:08: It's not like it is bad per se... It's about the mechanics of latency and fragmented context.

00:04:17: What do you mean by latency in this context?

00:04:19: Well, In a traditional stack your intent data from your website sits on one tool right.

00:04:23: Yeah Your CRM pipeline is another.

00:04:27: Your email engagement isn't.

00:04:28: third By time all those systems sync with an external lead scoring vendor hours or even days have

00:04:34: passed And the buyer has already moved on.

00:04:37: Exactly,

00:04:38: On a unified data foundation The internal lead scoring model Sees every single touch point instantly.

00:04:44: In real time It can trigger an action.

00:04:46: The precise second A prospect shows high intent Which

00:04:49: means you are catching the buyer While they're still in pricing page Not next day when reading lunch.

00:04:55: Exactly.

00:04:56: and it's not just Giant tech infrastructure companies Pulling this off either.

00:04:59: I saw that Megan Eisenberg The CMO at Samsara Is great example here.

00:05:03: Oh yeah her setup is incredible.

00:05:05: She is currently running twenty-five AI native tools and twelve custom agetic use cases right on top of their data layer.

00:05:13: Her philosophy is basically, hey if it's just about accessing data easily we don't need a vendor for that!

00:05:18: We're going to build it

00:05:19: ourselves.".

00:05:20: And this brings us back into the scenario you opened this deep dive with.

00:05:23: David V Kimball analyzed a hypertail of custom-built software.

00:05:31: Yes, he specifically talked about generative software.

00:05:34: Right the five minute software

00:05:35: Explain how that actually works in practice like day to day.

00:05:39: So let's say your marketing ops team needs to merge A highly specific list Of webinar attendees With historical lists of churned accounts To create a custom retargeting segment.

00:05:50: Okay super common request.

00:05:51: In the old days, you'd have to buy an expensive data orchestration tool or submit a IT ticket that takes like three weeks.

00:05:59: Or just give up and do it manually in Excel for ten hours

00:06:02: Exactly, but now an AI agent writes a custom Python script executes the data merge perfectly on your unified beta layer gives you this segment and then immediately deletes the script

00:06:13: because The software doesn't need to exist permanently.

00:06:15: It just needs to solve the immediate problem right there.

00:06:17: in their exact I mean if everyone is buying the exact same thirty commercial sauce products No one has a competitive edge.

00:06:25: The hypertail is where you build operational speed that your competitors literally cannot buy.

00:06:31: Right, it's custom to you

00:06:32: but Okay, if the data foundation is now this flat unified layer and we have custom software popping in an out of existence Who was actually running?

00:06:41: The day-to-day marketing.

00:06:42: well We are shifting from AI being a simple generate content button to AI becoming the actual orchestration layer.

00:06:50: Naveen Bonsall highlighted.

00:06:52: A recent Gartner prediction that I think it's pretty wild.

00:06:55: Oh

00:06:55: what do they predict?

00:06:56: Gartener

00:06:56: says that by twenty twenty eight sixty percent Of brands will be using agentic AI.

00:07:01: They are essentially signaling the death of traditional channel-based marketing.

00:07:04: Wait, by channel based marketing you mean having like a dedicated manager for LinkedIn ads?

00:07:09: A different manager for Google search and someone else running email automation

00:07:13: Exactly instead of humans manually adjusting bids an AB testing copy in isolated silos AI agents will manage the budget The creative iteration, the targeting and bidding simultaneously across all channels.

00:07:27: Just optimizing for a single revenue target?

00:07:30: Here's where it gets really interesting

00:07:31: It does!

00:07:32: Jorge Kuna took this step further.

00:07:34: He noted that AI agents are expected to orchestrate one trillion dollars in BDC revenue by twenty thirty.

00:07:40: Okay wait If I let an AI agent run my entire budget and dynamically change my creative across channels twenty-four seven without waiting for my Monday morning dashboard review, yeah.

00:07:52: How does a B to B team not lose total control of the brand?

00:07:55: Well that anxiety you're feeling right now... That is defining tension in this new era.

00:07:59: The

00:07:59: stress just thinking about it!

00:08:00: But major platforms recognize this and they are pivoting their entire architectures.

00:08:10: They are completely rebuilding it from the ground up specifically to handle the governance of that agentic AI

00:08:16: era.

00:08:16: And Katya Kisum noted a similar massive move by Adobe, right?

00:08:20: they just released an MCP server for Marketo and Gage.

00:08:24: Yes!

00:08:24: The Model Context Protocol Server.

00:08:26: This is very crucial development.

00:08:28: Okay let's translate this for the listener.

00:08:30: What does an NCP server actually do?

00:08:32: Is basically a secure translation layer between my private marketo data in an AI.

00:08:37: That is precisely what it is.

00:08:39: Because if you want an AI agent to analyze your Marketo data and suggest campaign optimizations, You can't just upload your proprietary customer list To a public model like ChatGPT.

00:08:50: No!

00:08:50: Your legal team would have a heart attack.

00:08:51: Exactly So.

00:08:52: an MCP server allows the model like Claude to securely query your MarketO instance Understand the context of your campaigns And take actions without ever absorbing your private Data into its public training set.

00:09:03: Oh wow Yeah, it basically turns Marketo into an open secure platform for custom AI agents.

00:09:09: You know if we are tearing down these walls and letting AI agents touch our core data securely my immediate thought is that this changes the entire dynamic of the buyer-seller relationship.

00:09:20: oh totally

00:09:21: because if vendors are opening up their systems to AI The buyers are doing the exact same thing on their end aren't they?

00:09:27: They

00:09:27: absolutely are!

00:09:28: And This might be the most challenging reality for marketers to accept right now.

00:09:32: The AI isn't just sitting on the vendor side of the table anymore.

00:09:36: Darnit Harvey Murray pointed this out brilliantly, he noted that B-to-B buyers are actively using tools like Claude, Perplexity and Chat GPT to explore categories compare providers read through documentation and build short lists.

00:09:51: And they're doing all these before a sales team even knows an opportunity exists.

00:09:55: Exactly Gartner.

00:09:56: actually projects by Ninety percent of B-to-B buying will be AI agent intermediated.

00:10:02: Wait, so if my Agenic AI is doing the marketing and their Agenetic AI Is Doing The Buying are we just setting up a giant bot to bot chat room?

00:10:10: Basically yes

00:10:11: But how does it b-tob brand actually prove its expertise through machine?

00:10:15: A Bot doesn't care If My Brand Video has high production value or great soundtrack.

00:10:19: And you've hit on the exact paradigm shift.

00:10:22: clean data is the prerequisite But the new battleground is a concept called algorithmic legibility.

00:10:27: Algorithmic Legibility?

00:10:28: Yeah,

00:10:29: for decades marketing has been about influencing human psychology through emotion color and narrative.

00:10:34: but an AI agent evaluates you based on structure logic and verifiable outcomes.

00:10:40: so if your brand's expertise in your customer Outcomes aren't structured in way that an AI Agent can read and trust You simply won't make this shortlist.

00:10:48: Exactly

00:10:49: So.

00:10:50: algorithmic Legibility is basically like SEO, but instead of optimizing a single blog post for Google's search crawler.

00:10:58: You are optimizing your entire company's operational reality for an AI buyers agent.

00:11:02: that Is the perfect way to think about it.

00:11:04: if you're pricing as obfuscated or If your case studies are logged in unstructured PDFs Or if your technical documentation is behind a lead capture form The buyer's AI cannot read and just

00:11:15: hit the wall right?

00:11:16: It will simply recommend the competitor whose data is fully legible and accessible.

00:11:21: That is terrifying.

00:11:22: If you aren't structured correctly, You are totally invisible to the primary research engine of your buyer.

00:11:27: Yeah But that brings up a massive issue because to optimize for algorithms and manage our own autonomous agents We need intense internal discipline.

00:11:37: How do we keep our own AI agents from going completely off-the-rails or leaking?

00:11:42: Our data.

00:11:43: well John Miller had a really great perspective on this.

00:11:46: he argues Traditional marketing automation platforms need to evolve into what he calls decisioning services.

00:11:52: Decisioning Services, What does that actually look like in practice?

00:11:56: It

00:11:56: acts as the ultimate governance layer.

00:11:58: so if you have an AI agent orchestrating campaigns it needs context.

00:12:02: The decision service holds the operational rules the email frequency limits the specific brand voice guidelines and all the compliance guardrails.

00:12:10: Okay So gives the AI boundaries

00:12:12: Exactly.

00:12:13: Because without that context, an AI agent might recognize that a specific very aggressive email subject line gets a forty percent open rate and just blast it to your entire enterprise account list.

00:12:25: Oh no!

00:12:26: Yeah It's technically correct optimization but is strategically disastrous error.

00:12:31: The decisioning service keeps the AI aligned with actual business reality.

00:12:35: I

00:12:36: love how Clark Barron illustrated current chaos around this.

00:12:39: He wrote this hilarious April Fools post on LinkedIn from the perspective of a company's MarTech stack, basically confessing its sins.

00:12:47: Oh

00:12:47: I saw that it was too real?

00:12:49: It was!

00:12:49: The Stack basically says hey you contracted with six vendors but i actually secretly invited twenty-five unauthorized vendors through various tag managers and by the way because you blindly clicked approve on the terms of service all your competitors who use me can now see your sale pipeline is written as satire

00:13:09: because it hits incredibly close to home.

00:13:11: The security implications of this autonomous era are just monumental.

00:13:16: Shashi Bellam Kanda highlighted the exact threat factor by analyzing Wiz's recent launch of specialized AI security

00:13:24: agents the cloud security company.

00:13:26: Right,

00:13:27: Wiz recognized that human security teams cannot move fast enough to govern AI so they deployed a red agent ,a green agent and blue agent.

00:13:35: Wait

00:13:35: what are those agents actually doing in practice?

00:13:37: Are they just monitoring server logs?

00:13:39: No!

00:13:40: They're entirely proactive.

00:13:41: The Red Agent actively probes your AI applications Just like a malicious hacker would.

00:13:46: It runs thousands of pumped injection attacks trying to break you own system Wow.

00:13:50: And when it finds a vulnerability writes the code to fix it.

00:13:56: Meanwhile, The Blue Agent investigates live threats on.

00:14:08: You can no longer afford to just be the creative in the room.

00:14:17: You have to act like an enterprise architect, you have to understand data security and honestly... ...you have to be a savvy internal politician To get any of this

00:14:27: implemented.".

00:14:27: You absolutely do because The biggest bottleneck to this entire transformation isn't the technology It's the organization itself.

00:14:35: Yogidawadwa articulated this beautifully when she brought up the concept Of the shadow veto.

00:14:41: Oh man, every marketing director knows the pain of The Shadow Vito.

00:14:45: It's brood!

00:14:45: Its when you finally get a yes and budget is approved by your CMO only to watch the initiative bleed-to-death over eight months in an IT security review

00:14:53: Precisely.

00:14:55: Yogita points out that getting leadership approval is only twenty percent off work.

00:14:58: The project almost never gets officially cancelled by leadership.

00:15:01: No they'd ever say no

00:15:03: Exactly.

00:15:03: instead it gets killed by a silent NO.

00:15:07: IT raises an eight week data privacy concern.

00:15:10: The data engineering team quietly deprioritizes your API integration because it's not on their sprint board.

00:15:17: Legal drags the feet of vendor compliance.

00:15:19: Right,

00:15:20: approval and actual enablement are two very different things

00:15:23: Very different.

00:15:24: So if

00:15:24: you're a marketing leader trying to build this composable AI driven architecture How do you actually overcome the shadow veto?

00:15:31: Well, by pre-wiring your internal blockers.

00:15:34: If IT legal or data team are seeing initiative for first time after CMO has approved it You have already lost.

00:15:40: Do

00:15:40: bring them in early.

00:15:41: Yes The complexity of managing these internal stakeholders alongside the collapsing tech stack is exactly why Ken Madsen argues that chief marketing technologist role now absolutely essential.

00:15:54: You need a dedicated executive whose sole job to architect this entire system and navigate the internal politics.

00:16:03: Okay, so we've covered the grand architecture.

00:16:05: The bot-to-bot buying journeys... ...the intense security governance and the internal politics.

00:16:11: But let's bring this back down to practical reality Because at the end of day you still have a pipeline target to hit This quarter.

00:16:19: How do best BDB teams actually drive revenue with these systems today?

00:16:23: It starts with a fundamental shift in how they measure success.

00:16:27: Franz Reimersma noted that we are moving into an era of attribution two point zero.

00:16:31: What

00:16:31: does it look like?

00:16:32: Well, We're finally shifting away from these petty dashboard turf wars over whether marketing or sales gets credit for lead.

00:16:40: Think

00:16:40: of this

00:16:41: I know right.

00:16:42: Instead attribution is evolving to decision support system focused purely on tracking overarching revenue and customer margin.

00:16:49: That's much healthier way.

00:16:51: And speaking of healthy approaches to operations, I love the hack that Phil Gamash and Jenna Kellner suggested.

00:16:57: Oh so

00:16:57: smart!

00:16:58: They advise marketers to completely stop complaining to executive leadership about tech debt because leadership tunes that phrase out.

00:17:06: it sounds like an IT problem right.

00:17:08: instead reframe it entirely to leadership as revenue readiness.

00:17:12: If you say, hey we need to clear this CRM sync issue so he can launch our targeted sales campaigns twice as fast and hit our Q-three revenue milestone suddenly leadership cares deeply about fixing the technical problem.

00:17:24: It is brilliant internal positioning.

00:17:27: And once you have that revenue readiness and a clean data architecture, the execution playbooks become incredibly powerful.

00:17:34: Bill Stothopoulos shared a fantastic highly practical example of this...

00:17:38: Oh!

00:17:38: The outbound system?

00:17:39: Yes

00:17:40: he broke down a six-step Outbound System That drove accompany from zero to three million dollars in annual recurring Revenue.

00:17:46: Walk

00:17:46: us through the mechanics of it.

00:17:47: how do they use AI to scale From Zero To Three Million?

00:17:50: So they didn't reinvent outbound.

00:17:52: They just supercharged the mechanics with AI and data.

00:17:55: first, they stacked intense signals that in.

00:17:57: Just look at who visited the website?

00:17:59: The combined web site visits with recent job changes on social media engagement.

00:18:03: Okay

00:18:04: then they routed all that raw data through cloud code to automatically run qualification logic.

00:18:10: He I teared every single account into tier one two or three instantly

00:18:14: which means the sales team isn't wasting hours researching unqualified leads

00:18:19: Exactly.

00:18:20: Tier one accounts received highly personalized multi-channel outreach, tier three accounts receive higher volume automated outreach.

00:18:28: that was still relevant.

00:18:29: and here is the kicker.

00:18:30: when a Tier One account replied A Slack alert fired instantly so an account executive could respond within minutes capitalizing on immediate engagement.

00:18:39: That's surgical execution.

00:18:41: But we also have to talk about inbound paid media Because Justin Rowe pointed out a hard truth about LinkedIn ads.

00:18:48: He cited Sam Dunning's strategy, noting that paid ads on LinkedIn work best purely as an amplification mechanism for already strong organic foundation.

00:18:57: Yeah you cannot just buy new audience with ads if haven't done the hardwork of building an organic presence first.

00:19:02: Exactly If your organic presence is empty, you're just paying a premium to make awkward ignored cold introductions at scale.

00:19:09: That organic work is the non-negotiable foundation.

00:19:13: Michelle Lieben provided an incredible breakdown of Lemlist's forty two million dollar ARR funnel.

00:19:18: that proves this exact point.

00:19:21: The lemless numbers are crazy

00:19:22: and really are.

00:19:23: ten percent Of their entire team posts organically on LinkedIn drives two million website visits.

00:19:32: I really want to unpack why that works, because it's not just about getting impressions.

00:19:36: if you think about the modern BDB buyer they are overwhelmed with automated cold emails and generic ads.

00:19:43: when ten percent of Lemless team is posting authentic organic content They're building human trust at scale.

00:19:50: And how do they capture?

00:19:51: That trust?

00:19:51: by removing all friction from those two million visits no credit card required, easy in-app booking and a perfectly smooth self serve process.

00:20:04: Because if Lemless put a massive lead capture form or you know mandatory forty five minute discovery call on the way The trust they built on LinkedIn would instantly decay.

00:20:12: Oh it'd be gone in a second!

00:20:14: The

00:20:14: frictionless trial captures the Trust immediately.

00:20:17: It is not about having most expensive ad budget Or the most complex software stack The teams that are winning, or the ones combining a unified data architecture with relentless disciplined execution playbooks.

00:20:29: Simplification, openness and operating discipline – That is formula for the third age of MarTech.

00:20:35: Which brings me to final thought I want leave you today.

00:20:38: We've talked extensively about this massive shift.

00:21:07: Wow.

00:21:08: Yeah, it raises a profound question about the rising premium on physical human connection

00:21:34: in.