Key Takeaways
- AI can accelerate ABM, demand gen and go-to-market execution, but only when teams ground it in clean data, account-level research and human oversight.
- B2B marketers drive stronger pipeline when they influence buyers before conversion and preserve the full context of engagement through nurture and sales follow-up.
Demand Gen Report’s B2BMX Summer Camp brought together top B2B marketing practitioners built around one goal: smarter programs that drive real pipeline results.
The event featured sessions from Stephanie McArthur of Demandbase, Josh Baez of NetLine, Hans Fischmann of AdRoll and Jason Gladu of Convertr, each delivering practical frameworks for applying artificial intelligence (AI) across ABM, demand generation, go-to-market execution and data integrity.
Across all four sessions, three lessons kept surfacing. AI amplifies what you feed it, so your inputs decide everything; clean data and account-level research separate real pipeline gains from expensive noise. Humans stay in charge of strategy while AI runs the execution, a division that protects your team from handing over control before the guardrails exist. And winning pipeline starts long before a buyer raises a hand, which means showing up early and carrying the full context of that engagement into every touchpoint that follows.
The AI-Native GTM Stack: How AI Orchestration and Connected Advertising Are Reshaping Demand Gen
Fischmann, vice president of product at AdRoll, took attendees inside the AI-native go-to-market stack, focusing on AI orchestration, connected advertising, the Model Context Protocol (MCP) and conversational AI. Fischmann’s central argument was when systems and channels communicate with each other, pipeline growth stops depending on how many hours a team can grind.
AdRoll’s MCP server, which extends campaign data, high-intent account signals and optimization capabilities directly into the AI tools marketers already use, served as the session’s proof point, cutting a 15-minute analysis workflow to under 30 seconds in a live demonstration.
Practitioners should treat orchestration as connective tissue, not another siloed tool, according to Fischmann. Connected advertising and AI-native workflows reduce the manual handoffs between systems, freeing teams to activate audiences and drive pipeline instead of assembling reports and switching between platforms.
“The marketers who win in the next five years won’t be the ones who mastered the most tools,” Fischmann said. “It’ll be the ones that spent less time operating tools and more time owning strategy and owning the growth conversation.”
AI in ABM: What Actually Works (And What’s Just Hype)
McArthur, senior ABM strategist at Demandbase, opened her session by naming the paralysis many B2B teams feel when confronting AI and then handing the room a way out. Her “AI value ladder” framework maps how AI progresses from generation and personalization through to analysis, recommendation and orchestration.
In cautioning against reaching for fully autonomous orchestration before building the human guardrails required to govern it, McArthur advocated instead for a balanced “human-AI continuum” in which centralized intelligence platforms handle the heavy data lifting while humans retain control of strategy and relationships. Rather than chasing “random acts of AI,” teams should compress research workflows from days to minutes, validate outputs against a strict human oversight ratio and prove data trustworthiness before climbing higher on the ladder.
“The AI is always going to assist, but the human is always going to own that strategy,” McArthur said.
From Form Fill to Follow-Up: Demand Gen’s Survival Guide to Win Buyers Before the Shortlist
Baez, director of demand generation at Netline, built his session around a number that reframes everything: 92% of B2B buyers build their vendor shortlist before they ever talk to sales. His session made the case that the shortlist isn’t formed at the moment of conversion but shaped far earlier, through the content buyers consume and the brands they encounter during research.
Demand gen teams, Baez argued, must pull two levers to influence buyers before they become visible: content that clarifies the buyer’s problem and coverage across earned, paid and owned channels. Before the form fill, marketing’s job is to help buyers understand their problem and recognize a brand as a credible option. After it, follow-up must restore the context of the original engagement, build familiarity through structured nurturing and let the buyer’s behavior determine the next step. When handing off to sales, the full story matters, not just a contact record.
“A form fill gives you a lead, but it doesn’t always give you the full story,” Baez said. “Buyers are forming opinions long before they talk to sales, and once they finally do engage, your follow-up will either build on that interest, or it’ll let it fade.”
The Data Integrity Playbook for the AI Era
As Convertr’s chief operating officer, Gladu is coused on data integrity, arguing it’s the missing link in AI-powered go-to-market execution. His opening statistics set the stakes: 94% of companies suspect their data is inaccurate, 30% of inbound data is invalid at the moment of capture and the fully loaded cost of a bad record is $117.
In an AI-driven system, Gladu stressed that bad data doesn’t sit quietly; it multiplies, scales poor decisions at machine speed and, once embedded in a model, can’t be removed without a retraining process that costs between $1 million and $10 million per run.
To validate data at the source, a “data integrity firewall” is needed to cover capture, standardization and activation prevents bad records from entering downstream systems in the first place. Real-time validation costs $1 per record to enforce; cleaning bad data in a CRM costs $10; allowing it to reach an AI model can make the error effectively irreversible.
“AI scales the cost of bad decisions the same exact way it scales the value of good ones,” Gladu said.
AI Amplifies What You Feed It, So the Inputs Matter Most
Every session eventually circled back to data quality and human judgment as the foundation of effective AI. McArthur and Gladu approached the principle from different angles but landed in the same place: AI’s output is only as trustworthy as the inputs behind it.
McArthur emphasized starting with account-level research to prove data quality before scaling any AI motion. “To avoid random acts of AI, you have to start with the low-hanging fruit, account-level research, and prove the data is trustworthy before you scale,” McArthur said..
Gladu made the case that bad data in an AI-driven system doesn’t just cost accuracy; it costs organizational trust in the entire stack. “”Data integrity is the missing link in AI-powered GTM. Bad data at scale doesn’t just cost you accuracy; it costs you trust in the whole system,” Gladu added.
Humans Stay in Control of Strategy
Automation handles execution, but the direction of what is trying to be accomplished must remain in human hands. McArthur and Fischmann both drew a clear line between what AI runs and what people own. Fischmann framed the shift in terms of where marketers spend their time: AI-native workflows eliminate the manual reporting and context-switching that consume the workweek, freeing teams to focus on strategy and growth.
“AI-native workflows let your team activate audiences faster and reduce the manual reporting, so people spend their time on strategy instead of stitching systems together,” Fischmann added.
McArthur warned that fully autonomous orchestration creates a “danger point” when guardrails aren’t in place. “Keep a human oversight ratio, something like 80% AI and 20% human, so you’re validating messaging accuracy before you ever hand over more control,” McArthur said.
Meet Buyers Early and Deliver Context, Not Just Contacts
Winning pipeline depends on showing up before a buyer raises a hand and then carrying the full story of that engagement forward into every subsequent touchpoint.
Baez made context his central argument, describing follow-up that fails to restore the original engagement as the fastest way to let genuine interest fade. “The shortlist is shaped much, much earlier than conversion. Marketing needs to help buyers understand the problem and recognize your brand as a credible option, and sales has to inherit the context, not just the contact record,” Baez said.
McArthur echoed the point through the lens of relationship-building: AI can personalize at scale, but humans build the authentic connections that close deals. “AI can handle the heavy lifting, but humans build the authentic relationships,” she said. “That’s the part you can’t automate away.”
