The future of AI in Marketo, with Niranjan Kumbi
Key Takeaways from Kapturall’s Session with Adobe Niranjan Kumbi, Adobe’s Director of Products for B2B Go-To-Market Orchestration, sat down with Youri Kuper and José Tarzian for forty minutes. These are the four segments worth rewinding to. The native Adobe Marketo Engage MCP server ships with more than 150 tools and is aimed at teams who want to bring their own model. Claude, GPT-4o, or whatever their organisation has standardised on and connect it directly to Marketo Engage data and actions. The tools that matter most for day-to-day operations are smart list construction and flow step sequencing: the two tasks that consume the most time and carry the highest risk of silent configuration error. Both are now callable from any MCP-compatible client. You describe the segmentation in plain language; the model builds the rules. What Kumbi said about the security architecture is the part worth rewinding specifically. He was not reassuring in the way that PR-coached executives are reassuring. He described network isolation and prompt-injection hardening as baseline requirements built in from the start and then explained why: Adobe’s assumption going in was that someone would attempt to manipulate a connected agent through crafted content in lead records. That is a realistic threat model stated plainly. If you are on the fence about whether to use purpose-built agents or connect your own LLM via MCP, this section gives you what you need to actually decide. “The assumption we designed for is that someone will try. The security is not a layer we added — it is how the server was built.” — N. Kumbi Key takeaways: Kumbi mentions AJO B2B Prime in a way that sounds incremental, a new edition, early access expanding in June and if you are not already tracking the AJO B2B situation it is easy to file it as a footnote. It is not a footnote. AJO B2B Prime connects directly to Adobe Marketo Engage data without requiring an Adobe Experience Platform or CDP implementation. That removes the one structural reason most Marketo-first organisations have had for not evaluating AJO B2B seriously. The CDP prerequisite was not a small barrier. It was a multi-quarter, significant-budget project that made AJO B2B functionally unavailable to the majority of the Marketo install base. Prime changes that. Teams that have been running lead-centric nurture for years and watching buying-group orchestration become the new standard now have a realistic path to get there without rebuilding their data layer first. This is the path. Rewind to this one if you have been deferring that evaluation. “This is the path for Marketo-first teams that want account journey orchestration without a platform rebuild.” — N. Kumbi Key takeaways: Youri asks Kumbi directly about headcount, whether AI reduces the need for ops practitioners. Kumbi does not deflect. He names the Jevons paradox explicitly: the historical pattern where efficiency gains in resource use increase total consumption rather than decrease it, because cheaper capacity unlocks more demand. His argument is that teams that can run campaigns faster will run more campaigns, not the same number with fewer people. This is persuasive, but worth examining from the ops side. The limiting factor on campaign volume has rarely been operator speed it has been the quality bar. Most organisations run fewer campaigns because their ops team is the only thing standing between good work and embarrassing work. If agents take over the execution layer without a corresponding investment in review and governance, the Jevons dynamic goes in a direction Kumbi did not address. Watch this section and form your own view. It is the question the industry needs to be debating right now, and this is one of the more honest versions of that conversation available in a public setting. “Teams that can run campaigns faster will run more campaigns, not the same number with fewer people.” — N. Kumbi Key takeaways: The session confirmed general availability dates but did not cover rollout mechanics Whether features activate automatically, whether there is an admin toggle, or whether existing custom integrations need to be reviewed before agents begin acting on the same data. One infrastructure point worth noting for high-volume accounts: Adobe is migrating Marketo Engage’s API layer from a private to a public cloud model, with improved concurrency and throughput as the stated outcome. No specific numbers were given, but teams running sync-heavy middleware or large-scale integrations that regularly approach API limits should track this migration separately from the AI feature rollout. “The June date is firm. What is not yet published is how the rollout is administered.” — Y. Kuper Key takeaways: What changes when Adobe brings agents into Adobe Marketo Engage
Rewind #1 — The MCP Server architecture explanation
Rewind #2 — The AJO B2B Prime announcement
Rewind #3 — The Jevons Paradox exchange
What to watch before June GA
SIGN UP FOR OUR NEWSLETTER
Juan Pablo García
Sales Manager at Kapturall
Sales professional with over three decades of experience in the tech industry. I specialize in building valuable partnerships with key players across several sectors. My strength lies in my deep understanding of customers' needs, a dedicated approach to service, and a knack for creating lasting relationships. I combine tried-and-true methods with the latest sales and marketing technologies, always focused on delivering value. I believe that every interaction, whether it results in a deal or not, is an opportunity to serve the customer and pave the way for future opportunities.
Related Content
