The dark funnel just got a lot darker, thanks to AI, and your traditional attribution models are officially broken. Go ask your customers.
📊 12 episodes across 6 podcasts
⏱ 379 minutes of intelligence analyzed
🎙 Featuring: Greg Kihlstrom, Kieran Flanagan, Kipp Bodnar
One Big Thing
The traditional attribution models you're relying on are actively misleading you, and AI is here to prove it. As marketers lean into AI for deeper insights, the chasm between what our current tech says and what qualitative data reveals is widening.
Drew Pinta (Director of Growth Data, Ramp) highlighted a critical flaw in traditional Multi-Touch Attribution (MTA) models during his discussion on The Dave Gerhardt Show (from Exit Five). Despite MTA models suggesting Meta was three times more effective than LinkedIn, an AI analysis of their Gong calls exposed an uncomfortable truth: LinkedIn was mentioned three times more frequently by customers as an influence. This isn't just a discrepancy; it's a fundamental misread of channel impact that directly affects budget allocation and strategic direction.
"We scan all our gong calls... LinkedIn was mentioned, I think, three times more than meta, even though all of our prior measurement using, like, an MTA model, had said Meta was like, three times more effective than LinkedIn."
— Drew Pinta, Director of Growth Data at Ramp on The Dave Gerhardt Show (from Exit Five)
This insight underscores what many demand gen leaders have been feeling: buyers aren't adhering to neat, trackable journeys. The rise of dark social, influencer channels, and now AI-driven insights are shining a harsh light on the limitations of last-touch, and even multi-touch, models. What you think is working, based on your dashboard, might be lagging far behind what's actually driving pipeline. The message is clear: your existing models are biased towards click-through, easily attributable channels, while valuable "view-through" channels like LinkedIn are being dramatically undervalued. This has massive implications for where your budget is going and the channels you might be underinvesting in.
The Rundown
① Your AI content is average because your context is average.
The problem with generic AI output isn't your prompts; it's the lack of a shared, deep foundational context layer for your AI, similar to Pixar's "Brain Trust" model. (Kieran Flanagan on Marketing Against The Grain)
→ The Play: Build four foundational "files" for your AI: Audience Delight Profile, Creator Style, Market Positioning Map, and Customer Journey Intelligence. This ensures distinct, on-brand AI output that continuously improves with data.
② Stop measuring creative marketing like direct response.
The fastest way to kill innovative marketing ideas is to force them into a direct response measurement framework; instead, measurement needs to adapt to the marketing tactic, not the other way around. (Drew Pinta on The Dave Gerhardt Show (from Exit Five))
→ The Play: For brand-building or experimental campaigns, use incrementality testing and event studies to quantify impact that traditional attribution misses. This allows for risk-taking and proving ROI for "darker" activities.
③ AI makes B-players better, not just A-players more efficient.
Contrary to popular belief, AI provides larger efficiency lifts for B and C players in execution roles, allowing them to streamline repeatable tasks, rather than only making A-players in management positions marginally better. (Neil Patel on Marketing School - Digital Marketing and Online Marketing Tips)
→ The Play: Don't just focus AI training on your top performers. Implement AI tools and training for your broader team, especially those in execution roles, to unlock significant productivity gains across the organization.
④ IKEA's €1 billion AI strategy wasn't about cost-cutting.
IKEA used AI chatbot data to identify a customer need for interior design, which led to reskilling existing employees for a new design consultancy that generated €1 billion in its first year—a surprising move away from typical AI-driven cost reduction. (Eric Siu on Marketing School - Digital Marketing and Online Marketing Tips)
→ The Play: Analyze your existing customer data (chat logs, support tickets, search queries) with AI to uncover unaddressed customer needs that could become new, profitable service offerings, not just efficiency plays.
⑤ Your design team wants a custom Claude skill.
Brain Labs built a custom Claude 'mega skill' that allows non-designers across the organization to self-serve on-brand marketing assets, drastically reducing the design team's workload on repetitive tasks and ensuring brand consistency. (Liz Spektor on The Dave Gerhardt Show (from Exit Five))
→ The Play: Invest in developing custom "organizational skills" for your internal AI tools (like Claude enterprise) that leverage brand guidelines and integrate with your existing tech stack (Figma, Notion) to empower self-service content creation.
Signal Board
🔥 Heating Up
• AI as an Operating Model: The conversation around AI is shifting from a tool-centric view to seeing it as a fundamental operating model reshaping how work gets done and teams are managed. (Michelle Cooper on The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX)
• AI for Ad Creative is Becoming Standard: Most ad creative and media buying is already leveraging AI today to scale production, save time, and reduce costs, with major platforms integrating AI deeply. (Caleb Kruse on Social Media Marketing Podcast)
• Underpriced Social Media Advertising (Snap, Pinterest, X): These platforms are delivering high ROI and lower costs compared to Meta and Google, making them prime targets for budget allocation. (Eric Siu on Marketing School - Digital Marketing and Online Marketing Tips)
👀 On Watch
• AI Transformation in Marketing: Organizations are rethinking entire business processes, with new KPIs focusing on customer outcomes over traditional efficiency, signifying a full business transformation, not just tech adoption. (Michelle Cooper on The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX)
• AI-generated Video for Ads 🆕: Tools like Cling 3.0 and Sora 2 are enabling multi-scene video creation and lifelike AI actors, accelerating content creation and accessibility. (Caleb Kruse on Social Media Marketing Podcast)
• AI Personalities (Faceless Videos) 🆕: AI-generated personalities on platforms like TikTok are achieving millions of views by creating engaging, contextually rich content, challenging traditional content creation models. (Caleb Kruse on Social Media Marketing Podcast)
🧊 Cooling Off
• Chamath's 'Brands Go to Zero' Argument: The idea that AI will make brands obsolete due to product cheapness/copying is being debunked; status and envy continue to drive consumer behavior, making brand a significant moat. (Neil Patel on Marketing School - Digital Marketing and Online Marketing Tips)
• Traditional Marketing Attribution Models: Relying solely on MTA or last-touch models is proving detrimental, as they often misrepresent channel effectiveness, particularly for "darker" or less directly trackable channels. (Drew Pinta on The Dave Gerhardt Show (from Exit Five))
The Debate
The core role and competitive advantage of marketers in an AI-powered world.
🐂 The Case For: Marketers are better positioned than engineers in the age of AI because the ability to understand customer intent, create compelling narratives, and execute creative strategies becomes paramount as AI handles the more technical, repeatable tasks. Marketing Against The Grain highlighted how foundational context (Audience Delight Profile, Creator Style) is the real differentiator, not just prompt engineering. This makes marketers the architects of AI output, guiding its strategic application. Drew Pinta also emphasized that the "fastest way to kill creative marketing ideas is to try to measure them like direct response," suggesting that sophisticated marketing judgment, not just technical prowess, is key in the age of complex measurement.
🐻 The Case Against: While AI automates tasks, it also poses a risk of commoditizing some marketing roles, shifting value to those who can build and manage complex AI agent workflows. Marketing School - Digital Marketing and Online Marketing Tips noted that the Anthropic team now manages multiple AI agents in parallel rather than writing code, suggesting marketers will need to become more like "product managers" for AI, emphasizing technical oversight and workflow building over traditional creative or strategic roles. Dave Gerhardt also raised concerns about AI agents eliminating human involvement in B2B buying, hinting at a future where even higher-level strategic marketing might be delegated to AI.
Our Read: The truth lies in the integration. Marketers who combine deep customer understanding and creative strategy with the ability to build and manage AI agent workflows will thrive. Those who cling to one without the other will struggle.
The Bottom Line
AI is forcing a demand gen reckoning: your data is lying, your buyers are hidden, and the only way to adapt is to build smarter context and continuously test.
Your Move
1. Audit attribution models: Don't just trust your MTA. Launch a quick survey or sales team interview to ask "What channel introduced you to us?" to compare against your marketing data (as seen on The Dave Gerhardt Show (from Exit Five)). You might find a significant discrepancy.
2. Build your AI's "Brain Trust": Start documenting your Audience Delight Profile, Creator Style, Market Positioning Map, and Customer Journey Intelligence in a central, accessible location. Feed these to your common AI tools as foundational context for better output (inspired by Marketing Against The Grain).
3. Pilot an "organizational skill" with AI: Identify a repetitive, on-brand creative task (e.g., slide deck generation, social media image design) that frequently pulls from your design team. Research and test building a custom AI skill (e.g., in Claude Enterprise) that enables self-service for non-designers, as detailed on The Dave Gerhardt Show (from Exit Five).
Channel Check
💬 Organic Channels
The dark funnel continues to deepen, with AI analysis of Gong calls revealing channels like LinkedIn driving significantly more influence than traditional MTA models credit (The Dave Gerhardt Show (from Exit Five)). This underscores the need to lean into qualitative data and direct customer feedback to understand organic discovery and influence points, rather than relying solely on last-click data.
📢 Paid Channels
While Meta and Google remain dominant, there's significant overlooked value in "underpriced" social channels like Snap, Pinterest, and X, offering higher ROI and lower costs (Marketing School - Digital Marketing and Online Marketing Tips). AI is also rapidly transforming ad creative production, with most ads anticipated to be AI-enhanced or fully AI-generated soon, across all major platforms (Social Media Marketing Podcast).
🧑💻 AI & Automation
AI is not just a tool; it's rapidly becoming an operating model, particularly in CX, where it's enabling frontline teams rather than replacing them (The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX). The emphasis is shifting from generative AI experimentation to operationalization at enterprise scale, requiring a redesign of creative and marketing workflows to achieve significant ROI and make B and C players more effective (The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX, Marketing School - Digital Marketing and Online Marketing Tips).
📖 Want the full episode breakdowns, guest details, and listen links?
Quick Appendix
The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX: "#840: Adobe's Hannah Elsakr on what happens after the hype: operationalizing AI at enterprise scale" · 37 min · Featuring Hannah Elsakr
Listen if: You're navigating the complexities of scaling AI beyond experimentation in an enterprise setting, and need practical advice on workflow transformation and responsible AI frameworks.
The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX: "#841: NiCE CMO Michelle Cooper on the most common mistake brands make with AI and CX" · 21 min · Featuring Michelle Cooper
Listen if: You're looking to redefine your AI strategy in customer experience, shifting from a technology-first approach to a 'customer moment-first' model that empowers frontline teams.
The Agile Brand with Greg Kihlström®: Expert Mode Marketing Technology, AI, & CX: "#842: Braze Chief Product Officer Kevin Wang on how AI has forever changed product development" · 28 min · Featuring Kevin Wang
Listen if: You're a product leader trying to balance rapid AI adoption with strategic decision-making that focuses on real customer value and fostering human creativity.
The Dave Gerhardt Show (from Exit Five): "Creative + AI examples from B2B marketers" · 56 min · Featuring Liz Spektor
Listen if: You want concrete examples and frameworks for how B2B marketers are actually using AI for creative, especially for self-service design and brand consistency.
The Dave Gerhardt Show (from Exit Five): "Inside Ramp's Marketing: Creative Bets, Measurement, and AI Agents (with Drew Pinta)" · 63 min · Featuring Drew Pinta
Listen if: You're questioning your attribution models and want to understand how a high-growth company like Ramp is tackling measurement for brand and performance, with AI insights from Gong calls.
Lab-Grown Marketing by Evidenza: "From Ad Skeptics to Ad Empires: Why Every Tech Giant Becomes an Ad Company" · 30 min · Featuring Jon Lombardo
Listen if: You need an elevated perspective on advertising as essential economic infrastructure and want to counter the internal narrative that ads are a necessary evil.
Marketing Against The Grain: "The AI Skill Ladder (Beginner → Workflow Builder)" · 19 min · Featuring Kevin Hutson
Listen if: You're a marketer looking for a clear framework to develop your AI skills, from novice to master workflow builder, using practical examples of AI agents and custom tools.
Marketing Against The Grain: "The Real Reason Your AI Content Is Average (It's Not Your Prompts)" · 24 min · Featuring Kieran Flanagan
Listen if: Your AI-generated content is generic, and you need to understand how to build a foundational context layer to produce distinctive, on-brand output beyond basic prompt engineering.
Marketing School - Digital Marketing and Online Marketing Tips: "Branding Just Changed Forever" · 20 min · Featuring Eric Siu
Listen if: You're challenging common narratives about brands becoming obsolete in an AI world and want to understand how status, envy, and the shift to "work per output" are redefining brand power.
Marketing School - Digital Marketing and Online Marketing Tips: "IKEA Just Found the AI Strategy Most Companies Missed" · 18 min · Featuring Eric Siu
Listen if: You want to see a tangible example of a company leveraging AI to create new revenue streams and services by reskilling employees, rather than just cutting costs.
Marketing School - Digital Marketing and Online Marketing Tips: "The BEST Social Media to Market On Right Now (It's Not Even Close)" · 20 min · Featuring Neil Patel
Listen if: You're looking for underpriced social media channels with high ROI and want to understand AI's nuanced impact on productivity across A, B, and C players.
Social Media Marketing Podcast: "Ads and AI: Leveraging AI Creative in 2026" · 43 min · Featuring Caleb Kruse
Listen if: You're planning your ad creative strategy for the next 1-2 years and need to understand the practical applications, ethical considerations, and future trends of AI-powered image and video generation.
