Marketers urged to clean data before automating customer journeys
Panelists at the MarTech Conference argued that automation speed must be balanced with buyer relevance, and that AI-driven engagement fails when underlying data is broken.
Marketers who push contacts down fixed automated tracks risk alienating the very buyers they hope to convert, according to a panel at the November 2026 MarTech Conference. Speakers argued that the industry's rush toward faster, more automated engagement has to be balanced against relevance, and that advanced decisioning tools cannot compensate for weak data.
The session, titled « From journeys to decisions », was moderated by Brianna Miller, demand generation director at Cohere Health. Panelists were Tim Boylan, Carlos Hidalgo, Mehul Nagrani and Jodie Wilson. Its central question was how to reconcile automation speed with buyer relevance in an era when generative tools are flooding digital channels with low-intent activity.
Nagrani compared enterprise buying to soccer rather than American football, saying that B2B buying committees debate, change their priorities and move in nonlinear ways rather than advancing through a predictable sequence of steps. In that environment, he said, a rigid journey built on a fixed set of stages is likely to miss the real state of a deal.
Hidalgo made a related point, arguing that customers do not buy in campaigns and that it is marketing that runs them. The implication, he suggested, is that companies should focus on what buyers actually need at each moment rather than on the cadence of their own programs.
The panelists drew a practical distinction between types of requests. Standard product specification requests, they said, should receive an immediate response, because speed is what the buyer wants. Nuanced relationship moments, by contrast, call for human judgment and time. Boylan argued that chatbots should be designed to recognize when they have reached their limits and hand the conversation off to a person.
Wilson described the division of labor between machines and people. Machine learning, she said, is good at surfacing topics, channels and timing, but it is the architects of a system who set the rules it operates under. Signals also need interpreting carefully: a chief financial officer downloading a document may point to budget or compliance concerns, while operations staff activity may indicate workflow pain. Conflicting actions from different parts of the same account may reflect two separate business units with different priorities rather than a single confused buyer.
The most pointed warning concerned data. Advanced decisioning, the panelists agreed, cannot repair a broken data foundation. Nagrani said AI dashboards can look convincing until the source data is audited, and Boylan offered an example from a rental company whose SMS opt-ins were routed to the wrong fields, undermining the automation built on top of them.
As a remedy, the panel recommended cleaning data around one or two lifecycle inflection points rather than attempting a complete overhaul at once. Nagrani suggested a diagnostic exercise: trace a single stalled opportunity through the system to see whether the organization is actually ready to act on its data. Hidalgo advised pairing technical audits with feedback from sales, customer success teams and buyers themselves. Recordings of the session are available on demand from the conference.
Source: martech.org
20
