- Organization issues as roadblocks to successful deployment (Chris Adelman and Lee Hammond)
- Buyer agents as increasing the importance of trusted information sources (Jon Miller, Dan Cote)
- Some vendors are already delivering end-to-end agentic campaigns (Matthew Niederberger, Duarte Garrido)
Thanks to everyone who contributed.
What I didn’t see was push-back against the vision that ChatGPT presented. Maybe people were being polite or maybe they fear that ChatGPT could hold a grudge. But my own initial take was triggered by words including “narrative” and “journey”. These led me to think that ChatGPT was still describing today's standard of multi-step marketing campaigns applied to customer segments. On closer examination, the story is a bit more complicated: the various components included in ChatGPT’s list are not necessarily a consistent whole, but rather a collection of predictions that ChatGPT has gathered by scraping of industry discussions. Oddly enough, ChatGPT's summary provides a clearer statement of an alternative approach:
AI-native marketing is:
- Continuous (not campaign-based)
- Conversational (not broadcast)
- Collaborative (AI agents on both sides)
- Generative (creating narratives, products, and experiences dynamically)
- Ethically-audited and explainable (trust is as important as persuasion)
Of course, that ChatGPT says something is no guarantee that it’s correct. So, although I intuitively agree with the prediction, I still need to convince myself (and, hopefully, others) that it’s right. One test is to compare the prediction with cutting-edge industry developments. Here’s a list of trends I’ve been tracking:
- Agentic search: AI-generated summaries replace traditional, link-based search results
- Ads in chat replies: Chat responses include paid advertising, identified with varying degrees of transparency
- Social commerce (a.k.a."live shopping"): Make purchases of items featured in social media discussions without leaving the social media platform
- Commerce via agentic browsers: Advertising and shopping links presented by browsers based on what they see the user doing (a privacy nightmare but we'll ignore that for now)
- Interactive ads: Advertisements that link to a conversational interface rather than conventional landing pages or websites
- Buyer agents: AI agents that do product research and, perhaps, make purchases on behalf of users without users visiting seller websites
- Conversational websites: websites that present chat-style interactions or hyper-personalized contents in dialog with the user
All but the last of these occur at the discovery stage of the buyer cycle. That's encouraging because marketers have been very concerned about they reach new customers when traditional search no longer drives traffic to their websites and consumers increasingly ignore traditional email and display messages. But what's really important is that these items all change the role of marketing from creating campaigns that broadcast messages to consumers, to executing continuous conversations with consumers. This is indeed what ChatGPT described. It’s not something that is accomplished by using AI generate conventional broadcast campaigns.
In fact, I’d argue that the current broadcast campaign model developed precisely because the then-existing technology couldn’t handle a conversational approach. Branching campaigns with personalized content nodes were the best that could be done when marketers were limited to deterministic, manually-crafted workflows. If you recall the once-popular cliché of marketing automation as a way to replicate the personal relationship of customers to their “corner grocer,” you quickly realize the real-life grocer had conversations, not campaigns.
Think a bit deeper and you see that the reason the grocer was able to have conversations was data. He knew the history of each customer and was able to discern their current situation by seeing and talking to them. Multi-step marketing funnels exist because marketers don’t have that kind of information available in real time. Instead, they have to make a guess at where each customer might be in the buying cycle. The funnel is a framework for making those guesses. Provide real information and it’s no longer needed.
I think this is important. Marketers, vendors and analysts have focused largely on how AI can reduce the cost and time of generating marketing campaigns. This allows companies to create near-infinite amounts of personalized content. But unless that content is driven by near-complete, real-time data, there’s a severe limit to how relevant it can be. Data, not content creation, is what ultimately limits the effectiveness of AI marketing.
This may strike you as bad news. Data has always been a problem for marketers, and real-time data is a particular challenge. Agentic campaign creation won’t solve the data problem, although perhaps it will highlight data's role as a bottleneck. The good news is that other AI and agent applications can in fact improve data access, something that data management vendors are already pursuing.
It’s true that better data management isn’t as exciting, or directly tied to revenue, as better customer engagement tools. This is probably why most vendors focus on adding engagement features. But as autonomous engagement become increasingly common, the need for better data will become more obvious. Vendors who use AI to deliver better data (and convince the market that they are leaders in doing it) will have an increasing advantage in growing their business.
I’ll give the final word to Abhi Yadav of iCustomer, one of those very smart people who see further, sooner than most of us. He recently shared a substack post that captures my opinion quite eloquently:
We used to drag customers to our properties (site, app, landing pages). Now we take capabilities to where customers already live: agent/app ecosystems, workflow-native GPTs, and search/ad surfaces that embed decisions at the edge. ... If your data stays scattered and your channels siloed while this shift accelerates, AI won’t transform your business it’ll just perform visual eye candy. AI should drive a paradigm shift in your business, not a party trick.
