Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

AI is everywhere, but marketing leaders face a harder question: Where should we actually use it?

In this episode of B2B Orchestration, Ian and AJ unpack a practical approach to bringing AI into a marketing organization. The conversation starts with the human side of adoption, including the natural anxiety people have about AI changing their jobs, and argues for positioning AI as a way to enhance and supplement the work teams already do.

From there, the discussion gets tactical: start with technology you already own, identify concrete use cases, create space for experimentation, establish appropriate governance, and focus investment on measurable outcomes instead of chasing every new AI tool. The conversation even uses Adobe Marketo as an example of looking for AI capabilities inside technology an organization is already paying for.

The episode ultimately explores the question every organization will face: when should you buy an AI capability, and when does it actually make sense to build?

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

Marketers aren’t lying to executives — they’re just explaining the end result without the how. That’s the distinction this conversation is built around, and it’s why multi-touch attribution keeps losing credibility with CFOs. The theory is elegant: distribute opportunity value across every touchpoint that contributed to a deal. But once you ask how the model actually assigns that credit, the story falls apart — it’s glorified last-touch dressed up as science, and the moment a touchpoint doesn’t get “the credit it deserves,” someone manually adjusts the model. Add in the fact that no one can closed-loop report an LLM citation, and you’ve got a measurement framework that’s been running on borrowed trust for years.

The conversation traces where that trust erodes and what’s replacing it. Instead of assigning dollar values to individual milestones — a blog visit, a badge scan, a demo — the shift is toward understanding causation: what actually moved someone to engage, and can you replicate it. This is where journey analytics and Markov modeling earn their keep — not by pricing out each interaction, but by modeling the path itself as the unit of value. That reframe matters more than it sounds: it’s the difference between measuring activity and measuring intent.

The second half applies the same lens to lead scoring. The 100-point threshold gets called out for what it is — a playbook that expired a decade ago, when scoring was never meant to declare someone “ready to buy.” It was designed to indicate engagement and content affinity, full stop. The episode walks through how that got corrupted — MQL, then AQL (automation qualified lead) as a stopgap after sales stopped trusting MQLs, then back to a form of MQL — and lands on where scoring is genuinely useful today: as a decay-based signal of fading intent, not a qualification gate.

The takeaway for anyone running attribution or lead scoring today: stop defending the model and start defending the story it’s supposed to tell. If you can’t explain how a number got generated, the number won’t survive contact with a skeptical buying committee — whether that committee is your CFO or your sales team.

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

This episode of the B2B Orchestration Podcast tackles a question most MarTech leaders are avoiding: if AI agents can already write, approve, and deploy campaign content for pennies, what happens to the traditional budget-and-roadmap model that’s run B2B marketing ops for a decade?

The conversation opens with a blunt read on disruption speed — the host points to Chegg and Quizlet as early casualties of AI eating search-dependent business models, then pivots to what that pattern means inside marketing operations. The guest, who runs an AI consulting practice, walks through a real deployment: an agent built on Claude (Sonnet for writing, Opus for technical work, Fable for planning) that takes a content brief, writes the copy, and pushes finished emails directly into Marketo or Adobe Journey Optimizer via API — collapsing what used to be a full role (or an outside agency engagement) into a workflow a marketer just approves and sends. The number that anchors the whole episode: a 10-email nurture sequence, once the template and brief exist, costs roughly $0.10 to generate.

That economic shift forces a second conversation — how do you justify budget for something that cheap, and how do you measure value when the unit economics don’t look anything like the old model? The guest argues this is actually an easier budget conversation, not a harder one: leaders don’t approve budget for vague transformation anymore, they approve budget for a visible, ownable outcome. “I can see this thing, I know what it’s going to do” beats a slide deck every time.

The back half goes further, into why the standard 90/180-day or 12-month transformation roadmap is becoming a dead pitch. With model capability shifting quarterly — new releases, new agentic tools, new price points — multi-year MarTech commitments are a hard sell to execs who genuinely don’t know what next quarter’s tooling will make possible. The episode closes on a speculative but pointed thread: whether GitHub becomes an “Etsy for skills,” where individual builders sell narrow AI workflows directly, and why enterprise buyers — compliance, security review, and vetting requirements being what they are — are structurally unlikely to ever buy that way, even as solopreneurs and small shops already are.

Takeaway for MarTech/RevOps leaders: stop selling multi-year transformation. Sell a visible, cheap, ownable outcome first, then use that proof point to open the budget conversation for what comes next.

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

This episode of the B2B Orchestration Podcast is a reality check for anyone still pitching marketing plans on vibes. The host sits down with Raja of GNW Consulting to unpack why the old playbook — "trust me, we'll estimate this much and hit it" — no longer survives contact with today's buying committees.

The core tension: a generational shift in how leadership evaluates marketing spend. Raja argues that newer executives, having grown up inside the "marketing bullshit" of inflated PowerPoint wins and vague Q-over-Q narratives, simply won't accept forecasts without hard justification. A 5-10% miss against forecast is tolerable; a 30-40% miss demands a real explanation. The days of throwing out a number and moving on are over — every dollar now needs a defensible line back to outcomes.

That shift reshapes how Raja's team engages clients. Rather than accepting a stated goal at face value, they reverse-engineer it: pulling apart what's actually driving current results, what's working, what isn't, and building a realistic target from that evidence — instead of inheriting an inflated number a leader is already boxed into defending.

The conversation's sharpest turn is on AI. Raja is blunt that most of what passes for "using AI" in marketing — email copywriting, subject lines, LinkedIn content — is low-value noise. The real unlock is AI applied to data analytics and reverse engineering: querying account and revenue data conversationally (he describes running ad hoc analysis in Slack instead of building reports in Salesforce) to get faster, cleaner answers about what's actually driving pipeline. That's the "left brain" use case he trusts.

The "right brain" example is more novel: Leo, a GEO (generative engine optimization) agent his team built for clients. It doesn't write content — Raja is firm that AI-written content is still weak — but it audits and edits existing content from a technical/schema standpoint so LLM crawlers can actually parse and cite it correctly. He describes clients going from zero AI-driven citations to several within a fairly short window, crediting the win to how the content is structured for machine readability, not to what was actually written.

The practical takeaway: stop treating AI as a blanket solution and start piecemealing it into the specific parts of the workflow where it measurably compounds — data interrogation and technical content structuring — rather than the parts (strategy, writing, judgment) that still need a human running the reverse-engineering.

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

Most leaders assume AI is closing the gap between what executives expect and what teams can actually deliver. This episode argues the opposite: that gap is widening — and there's data behind it.

The core tension: executives increasingly hear that a peer "did this with AI," ask an AI system if it can do the same thing, get a confident yes, and greenlight it — without understanding what actually sits between a working prototype and something safe to run in production. The conversation traces why that's happening now in a way it didn't with traditional build-vs-buy decisions, and why the gap is structural, not just a communication problem.

Key frameworks and arguments raised:

• The "chasm" between piloting and scaling. Everyone is experimenting with AI right now — the divide is between orgs stuck in pilot mode and the (few) actually capturing ROI. The guest argues the difference comes down to consistently underestimated work: governance, UAT for probabilistic systems, and defining ownership.
• Probabilistic vs. deterministic thinking is a genuinely new skill. Traditional QA and UAT were built for deterministic software. Testing an AI agent for edge cases requires a different mental model that most orgs — and most VPs — haven't built yet.
• Build vs. buy has quietly changed shape. It used to be: an exec hears what a peer's team did, tells their VP to replicate it, and the VP scopes real engineering work. Now an exec can validate the idea directly with an AI tool in minutes, which compresses the perceived distance between "idea" and "shipped" — while the actual distance (governance, tuning, edge cases) hasn't moved. That mismatch lands squarely on VPs, who are increasingly expected to operate like software leaders even when that isn't their background.
• "Launch" is being redefined. For traditional software, launch means done — shipped, GA, stable. For AI agents, the argument is that launch is closer to day zero: the agent will produce wrong outputs at first, and the real work — tuning, refining outcomes, building feedback loops — starts after launch, not before.
• Ownership is the unresolved question. If refining an agent's output is ongoing work, who owns it? Is it the go-to-market team using the tool, or a separate function tuning the model? The episode frames this as one of the most urgent org-design questions leaders aren't yet asking clearly enough — and one worth putting directly in front of an executive team: who owns this, what does "done" actually mean here, and how do we catch drift before it compounds.

If you're trying to figure out why your AI pilots aren't translating into scaled ROI, or how to talk to your exec team about what "launched" should actually mean, this one's worth the full listen.

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

B2B buying signals are compressing — and it's not because buyers care less, it's because they're doing more of the work somewhere you can't see. In this episode, the host and Jessica dig into how LLMs have changed the shape of the buying cycle: prospects are showing up to calls having already asked ChatGPT, Copilot, or Claude to compare your product against a competitor's. They're building their own battle cards before your sales team ever gets a chance to hand one over. The conversation traces what that does to the signals B2B marketers have always relied on — website visits, content downloads, form fills — and why those signals are getting both scarcer and less reliable as more research moves off-property. It also touches on why GEO and AEO (optimizing for how generative engines represent your brand) are becoming as important as traditional SEO.

The second half of the episode tackles a debate coming up constantly in B2B orgs right now: build vs. buy, specifically for AI-powered tools. Jessica makes the case that the "just build it ourselves" instinct badly underestimates what off-the-shelf software was actually doing under the hood. A traditional UI wasn't just a UI — it was a deterministic guardrail system: who can click what, who has access to which data, what actions are and aren't allowed. Replace that UI with a prompt-driven AI layer and you've traded a bounded, testable system for a probabilistic one that will always produce an output, even a wrong one, and will keep trying rather than fail cleanly like traditional software does (no clean "404" for AI).

That shift changes what QA and UAT even mean. Traditional software testing covers discrete, known use cases. Prompt-based AI has to hold up against unlimited phrasings of the same intent — "improve my campaign" said a hundred different ways — which means the real cost of building your own AI tooling isn't the initial build, it's the ongoing tuning, refining, and governance required to keep the model inside the boundaries your business actually needs. Jessica's central point: whether you build or buy, you still have to define your own governance — who approves what, what "QA'd" means at your company, what data access looks like — and skipping that step is where AI implementations go sideways. She closes with a sharp analogy comparing model drift mid-conversation to cutting a kid's hair during COVID lockdown: once it drifts a little off, there's no correcting your way back — you start over.

If you're evaluating build-vs-buy for any AI-powered MarTech capability, or trying to figure out why your funnel signals feel thinner than they used to, this one's worth the full watch.

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

This episode tackles a question every B2B marketer eventually runs into: now that we've automated away the friction of asking prospects for firmographic data, what are we actually doing with everything we've collected?

The conversation opens with a callback to the old days of progressive profiling — forms that asked visitors to self-report company size, job title, or function as a proxy for intent. The host and guest agree that era is largely over. Data enrichment tools now pull that firmographic information automatically, which solves the "explicit data" problem but surfaces a bigger one: teams enrich their leads and then stall out, because bypassing progressive profiling on forms didn't actually change how they go to market.

The guest makes a sharp point about the behavioral signals B2B teams routinely leave on the table. Enrichment tells you a visitor works at an enterprise company, but it doesn't tell you they've read five white papers on a specific topic — a much stronger intent signal than firmographic data alone. She argues B2B has a lot to learn from B2C's approach to personalization here: using behavior as a signal doesn't require identity. You don't need a visitor's name to know they're interested in a specific product line based on what they're engaging with.

From there, the discussion turns to why this so often breaks down in practice. The core issue isn't collection — it's sequencing. The guest argues you have to know how you intend to activate data before you collect it, not the other way around, especially in B2B where the data model itself is inherently more complex: one-to-many relationships between accounts and contacts, multiple buying-committee members per deal, and people who sit in the buying committee for more than one product simultaneously.

The most memorable framework of the episode is the "Michelin star ingredients" analogy: you can have every ingredient a five-star kitchen would want sitting in your fridge, but if you can't assemble and activate them into an actual dish — or in this case, a targeted, scaled campaign — the data was never the point. The episode closes on a warning relevant to the current AI moment: having a mountain of enriched data and then defaulting to "just activate it with AI" (whether that's email generation or something else) isn't a strategy either. The real differentiator is a platform that actually understands the market well enough to recommend the next best action — not just execute on command.

Key takeaway: Enrichment gives you facts about a company. Behavior gives you intent. The teams winning in B2B right now are the ones building a data strategy backward from activation — not collecting first and figuring out activation later.

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

Every B2B marketer right now is hyped about MCP servers — and Drew's first move in this conversation is to puncture that hype just enough to be useful. His core argument: an MCP server isn't the destination, it's the on-ramp. The real value only shows up once teams get past the novelty and start building something with it. He draws a sharp comparison to the early days of APIs — back then, APIs lived entirely in IT's world. MCP is different because it sits at the intersection of development and marketing, which means the average marketer is going to need to understand what it does and doesn't do, not just consume it secondhand from IT.

That leads into the bigger structural question: what happens to the marketing function itself? Drew pushes back on the idea that marketing and IT simply merge into one job. Brand people, creatives, and social/media specialists aren't going anywhere — they get enhanced by technology, not replaced by it. But marketing ops and rev ops are a different story. Drew's blunt take is that the ops people who win the next few years are the ones building, troubleshooting, and maintaining AI agents themselves — which means working directly with CISOs and CTOs, touching infrastructure like Neon databases and Ingest, and yes, having a GitHub account with a real portfolio of shipped agents. His advice to anyone in ops: start that GitHub now, before you need it.

On hiring, Drew's answer is deliberately role-dependent rather than a blanket rule. Brand hires still get evaluated on portfolio work, not code. But for ops roles, the bar is moving fast — within 6-12 months, hiring managers should expect to see a working portfolio of agents built, plus credentials (Anthropic, OpenAI, and the wave of university programs now teaching agentic AI workflows). The signal Drew's watching for: candidates are no longer coming in with a "marketing degree" as the qualifying credential — they're coming in with agentic AI training regardless of their original field.

The most substantive part of the conversation is the build-vs-buy shift. Drew's thesis: organizations now have the ability to solve micro-problems themselves with agentic workflows built on tools like Claude or ChatGPT, instead of defaulting to "is there software for this?" That's not the same as building your own CRM or trying to replicate Marketo in-house — it's about right-sizing the problem to the solution. Counterintuitively, Drew argues this is currently slowing buying cycles, not speeding them up, because teams are over-indexing on "we can probably build this" before they've felt the real cost, which never shows up in the first two months — it shows up in maintenance, integration, and evolution six months later. He predicts a wave of public "I vibe-coded my own CRM and it wrecked my business" stories on LinkedIn and Reddit before the market recalibrates and organizations find real balance between build and buy. He closes with the Jurassic Park line — teams were so preoccupied with whether they could that they never asked whether they should — as the operating principle B2B leaders need right now.