Best of LinkedIn: Go-to-Market CW 12/ 13

00:00:00: This episode is provided by Thomas Hallgeier and Frenis based on the most relevant LinkedIn posts about go-to market in calendar weeks, twelve and thirteen.

00:00:08: Frenes is a B to B Market Research Partner helping ICT and tech providers identify niche channel partners by compressing the full journey from identification to qualified first meeting into four or five weeks.

00:00:19: you can find more info at the description.

00:00:21: all right.

00:00:21: so let's get into it.

00:00:22: yeah welcome everyone to The Deep Dive.

00:00:24: Yeah.

00:00:25: So imagine a marketing department that well Never sleeps, runs entirely inside a handful of command-line terminal windows and just earned its human architect a fifty thousand dollar salary bump.

00:00:36: Which is just wild to think about!

00:00:39: Right

00:00:40: I mean if you are building a BtoB revenue team right now the sheer volume of marketing tech hype your being bombarded with is likely deafening.

00:00:48: So we've been analyzing the absolute top go-to market trends across LinkedIn over the past two weeks, specifically calendar week's twelve and thirteen.

00:00:55: And there is a lot of noise out there right now

00:00:57: exactly.

00:00:58: so our mission for this deep dive Is to cut entirely through that noise for you.

00:01:03: We are looking at the fundamental architectural rewiring Of the twenty twenty six revenue engine

00:01:10: Yeah, and the trajectory we are seeing in these discussions is incredibly sharp.

00:01:14: I mean... We're moving far beyond the era of AI as just a simple writing assistant

00:01:19: Like a glorified spell checker.

00:01:20: Right!

00:01:20: Exactly.

00:01:21: We were looking at landscape where autonomous agents actually act like an execution engine And that requires totally new technical discipline What the industry now calling GTM engineering Which

00:01:32: will definitely get into?

00:01:33: Oh

00:01:33: absolutely.

00:01:34: But perhaps most critical thread today Is how plugging this hyper efficient system into, you know legacy commercial models is exposing catastrophic cracks in how modern companies operate.

00:01:46: So

00:01:47: let's actually start with that shift and the AI deployment model.

00:01:50: because we used to treat artificial intelligence like a highly caffeinated intern.

00:01:54: You had to micromanage or like a calculator

00:01:56: punching a prompt.

00:01:57: get an output

00:01:58: right?

00:01:58: You got a single output And then you have to manually move that out.

00:02:01: put into your workflow

00:02:02: yeah.

00:02:03: but Daniel Sulsarini points out in his recent analysis that AI is now operating much more like a thermostat.

00:02:10: A thermostat?

00:02:10: I like that analogy!

00:02:12: Yeah,

00:02:12: you set the desired state—like your revenue target or your pipeline quota and agenic.

00:02:18: GTM systems automatically turn various APIs scraping tools and outreach channels on-and off from background to achieve.

00:02:26: So it's an entire operating system, basically.

00:02:28: Exactly

00:02:29: not just a cold email generator.

00:02:31: And that distinction between you know A tool you log into and in environment you orchestrate is central here.

00:02:37: Joyce G actually provided a phenomenal look under the hood of what this looks like In practice.

00:02:42: Oh

00:02:42: yeah The terminal windows post Right.

00:02:44: She didn't ask LLM to write blog posts.

00:02:46: she built four parallel workflows turning clod code Into comprehensive GTM execution layer.

00:02:52: Wait hold on Four work flows in parallel.

00:02:54: Yes

00:02:54: She's literally running three to five terminal windows simultaneously.

00:02:58: I mean, as someone who works in go-to market strategy the idea of running command line interfaces carifies me.

00:03:04: is this realistic for standard revenue teams or Is it just like a parlor trick?

00:03:08: For highly technical founders?

00:03:10: well It's becoming the standard because of the sheer leverage it provides.

00:03:14: what Joyce is doing those terminals is orchestrating parallel autonomous sessions.

00:03:19: So one AI agent is querying search engines to scrape page-one rankings and identify keyword gaps.

00:03:25: Okay,

00:03:25: all in the background

00:03:27: Right.

00:03:27: And at the exact same time, in a separate terminal another agent is drafting a fifteen hundred word piece based on yesterday's data.

00:03:35: meanwhile A third Agent Is pushing an approved piece live via a CMS API and

00:03:40: pulling analytics right

00:03:41: exactly Pulling Google search console analytics to measure The impact.

00:03:45: the human isn't writing or publishing.

00:03:47: the Human is simply reviewing the Terminal outputs In you know authorizing the next operational loop

00:03:53: The compression of time there is what completely changes the unit economics of a marketing team.

00:03:57: I mean, Ho-Sun Chung shared a workflow that essentially algorithmizes competitive intelligence in a similar way.

00:04:04: Oh!

00:04:04: The battle card setup.

00:04:05: That was impressive

00:04:07: Right?

00:04:07: Instead an analyst spending days building a battlecard Chun's AI agent pulls realtime data from G to RFs built with an Owler simultaneously.

00:04:17: It just scripts the whole software stack?

00:04:19: Exactly, it parses their software stacks identifies there backlink vulnerabilities and formats a usable sales battle card in like fifteen minutes.

00:04:30: but if we're talking about complex multi-layered operations Brittany Duffy setup is the absolute extreme end of this spectrum.

00:04:37: Oh

00:04:37: for sure.

00:04:38: she builds us synthetic organization.

00:04:40: yes the multi-agent database architecture.

00:04:42: This is where we move from just automation to true autonomous operations.

00:04:47: Precisely She structured a system commanded by chief of staff agent.

00:04:50: So as human operator you only interact with Chief of Staff.

00:04:53: You give it high level strategy

00:04:55: Right, you feed at high level strategic directive and that central agent parses your intent and delegates tasks to four specialized scoped domain agents.

00:05:04: so there's one strictly for market research One for OX One for design And one for reliability testing.

00:05:09: And the fascinating part is the mechanics of how they talk to each other.

00:05:12: Yes, The Central Succulate Database.

00:05:14: All right?

00:05:15: They all share this database.

00:05:16: so...the market research agent populates a table with ICP data and the design agent automatically detects that new data, it begins generating assets tailored to those specific thermographics.

00:05:27: Without the human ever bridging the gap?

00:05:29: That is insane!

00:05:30: And if we look at the downstream impact of that architecture you know... It fundamentally alters the human sales floor.

00:05:36: Frank Sondors is applying this closed-loop system To the unglamorous highly manual work that SDR's notoriously neglect.

00:05:44: Ah

00:05:45: yeah, the stalled MQLS.

00:05:46: Exactly

00:05:47: He deployed a voice AI system.

00:05:49: he calls Agent Frank to tackle stalled inbound MQLs, missed demos and churned accounts.

00:05:55: So an actual voice agent calling people?

00:05:57: Yes!

00:05:57: The VoiceAgent handles the real-time dynamic conversation.

00:06:00: it parses the prospect's objections updates the CRM via API And then either rebooks meeting or disqualifies lead.

00:06:08: so the AI handles entire flywheel of execution

00:06:11: Exactly.

00:06:12: And the humans are left to do what humans actually do best, which is handling complex negotiations and building strategic relationships

00:06:20: Which brings us into human capital side of this whole equation.

00:06:23: If execution layer now lives in terminal windows Shared databases and API webhooks Standard marketing ops professionals simply don't have technical chops.

00:06:34: No they dont Which explains the explosion of gtm.

00:06:37: engineer role we've been tracking

00:06:40: But I have to push back a little on the longevity of this trend.

00:06:43: We're seeing Margeau-Voget and Matt Firestone report that technical GTM engineers are pulling in median salaries between one hundred thirty five thousand, and one hundred fifty thousand dollars.

00:06:53: Yeah which is massive fifty thousand dollar premium over traditional marketing ops.

00:06:57: Right but isn't that massive salary bump just temporary tax early adoption?

00:07:01: Like once these AI tools roll out user friendly drag & drop interfaces next year won't engineer title become obsolete?

00:07:09: Well, that is a really common assumption but the underlying architectural complexity suggests otherwise.

00:07:15: I mean... A drag and drop UI can help you build a simple email sequence?

00:07:18: Sure!

00:07:19: But enterprise routing requires secure infrastructure, complex data enrichment waterfalls and custom API connections that don't just break when a platform updates its endpoints.

00:07:30: So it requires a true engineering mindset?

00:07:33: Exactly!

00:07:33: Yeah And Uri Orlov actually contextualized this brilliantly by digging into organizational theory from the nineteen seventies.

00:07:41: On

00:07:41: those boundary spanning theory I am always down for applying nineteen seventys organizational theory to twenty-twenty six tech stacks.

00:07:47: Right It's so relevant.

00:07:49: So the theory dictates that as internal departments within a company become highly specialized, they start using their own unique jargon metrics and tools.

00:07:57: And a massive information asymmetry develops between them?

00:08:00: Exactly!

00:08:00: They essentially lose the ability to communicate In the seventies this required specialized human boundary spanners To translate between say R&D in manufacturing Okay...and

00:08:11: today

00:08:11: Today, the revenue function has gone entirely digital which creates a massive complexity gap between commercial strategy and technical execution.

00:08:19: Because

00:08:19: sales marketing in product are specialized silos operating on totally different data models.

00:08:25: Yes,

00:08:26: so Yuri argues that GTM engineering isn't some vendor invented buzzword to sell software.

00:08:32: it is an organizational inevitability.

00:08:35: You need a highly technical architect To build the bridges that allow data to flow between those silos.

00:08:40: Exactly

00:08:40: That reframes the role entirely and aligns closely with Accentune to Opalize argument that The title should actually be sales architect.

00:08:48: right because the AI as the engineer now

00:08:50: exactly The AI writes the Python scripts, generates the API calls and executes data transformations.

00:08:57: The human is the architect sitting above code designing the blueprint

00:09:01: And we are seeing that exact split between design & execution play out in tooling stack right now.

00:09:07: Jacob Tuener highlighted emerging dynamic between platforms like Claude Code and Clay.

00:09:11: Oh yeah!

00:09:12: Claude as a Design Layer

00:09:13: Right...Claude is functioning as a design and logic layer.

00:09:17: You talk to it describe your strategic intent, and it generates the programmatic logic.

00:09:23: But Clay remains the enterprise execution layer?

00:09:25: Exactly!

00:09:26: You still need that robust, secure infrastructure to handle complex data enrichment without hallucinating or you know breaking data privacy compliance.

00:09:36: And what's remarkable is this architectural revolution isn't geographically centralized in Silicon Valley anymore.

00:09:42: Tim Cardin and Fernando C have been documenting how Europe —and particularly France—is quietly building a massive GTM software empire.

00:09:51: Yeah the numbers out of Paris are wild!

00:09:53: Paris alone has twelve major GTM tools And across Europe, there are thirty-one platforms aggressively challenging US dominance.

00:10:00: We're talking about companies like Lemlist, Folk and LaGrowth Machine pushing the boundaries of autonomous

00:10:06: outreach.".

00:10:06: And you know with that explosion of tooling... The immediate assumption is marketing budgets must be spiraling out of control to afford all these platforms?

00:10:13: Yeah

00:10:13: for sure!

00:10:14: But the mark a true GTM engineer actually technical frugality.

00:10:19: Jorge B. Macias made a compelling point that you can actually gauge an architect's seniority by their software budget.

00:10:26: Wait, really?

00:10:27: By how little they spend...

00:10:28: Yes!

00:10:29: A disciplined architect keeps the entire stack costs between just five hundred and a thousand dollars per month.

00:10:35: That is incredibly low for an enterprise engine

00:10:38: Because they don't just buy out of the box size subscriptions, They utilize APIs efficiently.

00:10:43: Cache data to prevent redundant API calls and deeply understand token optimization when prompting LLMs.

00:10:49: Right but okay if API calls are that cheap And these agents or that efficient?

00:10:55: What happens when you plug a hyper-efficient execution engine into a fundamentally broken commercial strategy?

00:11:01: Well, that is exactly where the wheels are falling off for a lot of companies right now.

00:11:04: Yeah Eddie Reynolds has been very vocal about what he calls GTM FOMO.

00:11:08: Leadership teams just feel the pressure to adopt AI so they scramble to layer autonomous agents on top of their existing processes Which

00:11:15: just amplifies their current dysfunctions at light speed.

00:11:18: Exactly and Adam Jay echoed that exact sentiment.

00:11:21: He stated plainly your GTM isn't broken.

00:11:24: It was never built right in the first place.

00:11:25: You

00:11:26: cannot scale a foundation that was never engineered to scale!

00:11:29: Right,

00:11:30: if your product positioning is weak or you're sales team misaligned with marketing messaging adding multi-agent AI system just scales the noise.

00:11:39: it automates generation of highly personalized incredibly articulate garbage.

00:11:44: Which brings us to the concept of semantic data layers and what Hilaterbrack calls context engineering.

00:11:51: Context Engineering,

00:11:52: right?

00:11:52: It's one thing to say you know AI needs context to avoid hallucinating.

00:11:56: but the technical reality of providing that context is where teams fail

00:12:00: because if marketing needs The General Manager's approval to publish a blog But sales reps are outpitching their own rogue version on the roadmap there's no shared reality.

00:12:08: Exactly, so Context Engineering is the discipline of forcing the entire organization to align on single mathematically verifiable ideal customer profile and structuring that data so an LLM can actually comprehend it.

00:12:21: And the mechanics for how you structure this data are critical.

00:12:25: Doe Wester advocates a Forcing Function.

00:12:27: he calls The OnePage GTMOS.

00:12:30: Oh I loved that post!

00:12:32: Abandoning the gut feelings & vibes.

00:12:34: Right start with evidence not Vibes.

00:12:37: You run an analysis on the top twenty percent of your current customer base looking strictly at revenue retained and lowest cost of acquisition.

00:12:44: You map the actual buyer journey based on deals that mathematically closed?

00:12:49: Yes, not.

00:12:49: the theoretical funnel you're marketing team drew in a whiteboard!

00:12:53: You distill the ICP—the verified value prop —and the proven activation channels onto a single page because if your core strategy requires a forty-slide deck to explain Your agents will fail to execute it.

00:13:04: But then, once you have that crystallized strategy... You face the technical hurdle of feeding it to AI.

00:13:10: Right!

00:13:11: You can't just dump your entire product documentation repository into an LMM context window and expect flawless execution.

00:13:17: No definitely not.

00:13:19: Lorenz Nice issued a fantastic technical warning about phenomenon called Context Rot

00:13:23: context rot.

00:13:25: I mean, that is a brilliant way to describe the degradation of LLM performance when it's overwhelmed with

00:13:30: data."

00:13:31: It really is!

00:13:34: I mean a human can open at two hundred page manual, scan the table of contents and skip straight to page one forty in seconds.

00:13:43: But an AI does not read like that?

00:13:44: Exactly!

00:13:45: When you feed an AI massive data set it holds all those tokens into its active memory.

00:13:51: The more noise you introduce...the higher probability loses.

00:13:54: signal begins hallucinating.

00:13:56: connections don't exist.

00:13:57: So how do prevent context rot while still giving agent enough data?

00:14:03: Lawrence advocates for a system design called progressive disclosure.

00:14:07: Instead of dumping raw data into the prompt, you build a semantic architecture.

00:14:12: So you provide an agent with top-level navigation hub?

00:14:14: Right!

00:14:14: Essentially a summary what information exists and where to find it.

00:14:18: The agent is prompted evaluate specific task, consult the Hub then query vector database only for the specific granular deep dive chunks.

00:14:26: at that exact millisecond

00:14:28: It retrieves context, executes tasks and flushes its memory

00:14:33: Exactly.

00:14:33: Preserving its token limits for the next action.

00:14:36: That level of architectural discipline is staggering.

00:14:39: It completely flips the script on what we consider AI implementation.

00:14:43: We aren't really implementing AI, we are fundamentally cleaning up our own data so that AI can function.

00:14:49: Yes And Brendan Short recently published a study of eleven growth stage companies that successfully built internal GTM agents and his findings support this entirely.

00:15:00: Yeah, the dirty secret of these highly advanced revenue teams.

00:15:02: Right?

00:15:03: Exactly!

00:15:03: The actual work wasn't advanced prompt engineering...the heavy lifting was basic data hygiene

00:15:09: Which makes total sense.

00:15:10: If your CRM is riddled with duplicate accounts and you're third-party enrichment data is stale You are entirely unready for agentic deployment.

00:15:19: And Brendan noted something fascinating about how these teams generate requirements documentation For their AI agents.

00:15:24: They don't write complex technical specs.

00:15:27: No..The requirement doc is simply shadow your best rep.

00:15:31: Shadowing the rep to extract the disqualification signals?

00:15:34: It's so elegant!

00:15:35: You don't ask the AI to invent a new sales motion, you monitor your top-performing account executive

00:15:42: Right...you document the exact data points that make them lean into a deal and specific red flags that makes them disqualify a lead.

00:15:49: within five minutes.

00:15:50: You encode those human heuristics in the agent logic layer.

00:15:54: So the AI simply scales proven human intuition at machine pace.

00:15:58: And that dynamic between the technology scaling, The Human Intuition highlights a structural shift in how we think about competitive motes.

00:16:06: Oh this brings up to Gabriel Goy analogy.

00:16:09: Yes!

00:16:09: The analogy regarding satellite broadband perfectly captures it.

00:16:13: He compared Elon Musk's Starlink to Amazon project Kuiper or Amazon Leo.

00:16:18: Right, so Starlink has thousands of satellites in orbit.

00:16:21: It's massive bleeding-edge infrastructure.

00:16:24: But Amazon has a fraction of that.

00:16:25: So if you view the market purely through the lens of technology.

00:16:28: starlink possesses The ultimate advantage

00:16:30: but amazon possess is the ecosystem.

00:16:32: Yeah because starlink Is ultimately just selling the pipe the raw internet connection.

00:16:37: Exactly every satellite they launch has to mathematically pay for itself Through subscription revenue?

00:16:42: But amazon on the other hand views the pipe as a loss leader.

00:16:46: They use the satellite connection as a cheap customer acquisition channel to pull rural or disconnected users directly into the highly profitable Amazon ecosystem like Prime, AWS smart home integration.

00:17:00: So that technology stack itself is becoming a commodity?

00:17:03: The

00:17:03: infrastructure's the commodity.

00:17:05: go-to market Is the moat right

00:17:07: in our world.

00:17:08: the LLMs the scraping API is the parallel terminal workflows.

00:17:12: those are just the pipe.

00:17:13: Every company will have access to the exact same AI capabilities.

00:17:16: So, The true differentiator is your deep nuanced understanding of customers' pain points and proprietary ecosystem value.

00:17:24: you build around

00:17:25: them?

00:17:25: Exactly!

00:17:26: And this shifts the entire paradigm of organizational design.

00:17:29: If technology as a commodity that handles execution how do you structure human team around it?

00:17:34: Well,

00:17:34: Stephane has argued if you are rebuilding a go-to market system today You must completely stop mapping job

00:17:40: roles.

00:17:40: Yes You don't draw an org chart with an SDR, a content marketer and account executive.

00:17:46: You map task categories

00:17:48: Because AI does not replace the job title it obliterates a task category.

00:17:53: That is crucial distinction.

00:17:54: Instead of hiring traditional SDR layer to grind out cold calls Your AI system handles the task category of prospecting.

00:18:02: The agents monitor the internet for intense signals, cross-reference that against your one page GTMOS and draft highly contextualized outreach.

00:18:10: And the human account executive only steps into a loop when a prospect replies and demonstrates intent.

00:18:20: allows hyper lean teams to operate with the pipeline generation capacity of an enterprise organization.

00:18:26: And we saw the ultimate proof of this leverage in Mark Hardy's recent breakdown

00:18:30: Of a series B race.

00:18:31: Yes, he detailed how a tiny small but mighty GTM ops team utilize platforms like clay and clawed code To operate like a team five times its physical size.

00:18:41: by mapping tasks instead of roles and keeping their data ruthlessly clean.

00:18:45: Exactly, this micro team doubled their outbound sourced pipeline and successfully powered an eighty million dollar Series B raise.

00:18:54: An eighty-million dollar raise entirely driven by a fractional team armed with autonomous agents and pristine data hygiene?

00:19:03: That is just incredible!

00:19:04: It's the new baseline for b to be performance.

00:19:07: but you know it also presents a looming existential challenge for revenue leaders that we have yet to fully grapple.

00:19:15: Well, we are celebrating the fact that AI is commoditizing execution.

00:19:19: Any rep at any company can now spring up an agent to generate perfectly researched flawlessly personalized outreach in seconds.

00:19:27: but We're ignoring The other side of the equation

00:19:30: buyers.

00:19:31: if marketers or building hyper efficient agents to send perfect Outreach buyers will inevitably deploy their own AI gatekeepers to filter it out.

00:19:38: Wow

00:19:39: yeah

00:19:39: We are rapidly approaching a reality where our meticulously crafted go-to market bots are exclusively pitching to enterprise procurement

00:19:46: bots.

00:19:46: So, completely synthetic top of funnel?

00:19:48: Exactly!

00:19:49: When both sides possess infinite operational leverage traditional outbound marketing cancels itself out entirely.

00:19:56: If the tech is perfectly equal on both side's transaction The only thing that will bypass an AI gatekeeper Is genuine unfakeable human trust

00:20:05: and proprietary data that the LMS haven't already trained on.

00:20:08: Exactly!

00:20:08: That is a real challenge for twenty-twenty six.

00:20:11: Lots

00:20:11: negotiating with bots, leaving humans to only deal with final handshake…that's a wild reality.

00:20:29: Thank you so much for joining us on this deep dive into the architecture of modern go-to market.

00:20:34: Don't forget to subscribe, take a hard look at your data foundations and we will catch.