Personalisation's Real Problem Isn't Tech. It's Trust
As buyers delay contacting vendors until deep into their decision-making, the real constraint on personalisation isn't technology — it's trust.
Real personalisation means meeting people where they are with what they want. Delivering on that requires infrastructure that can genuinely listen at scale, replicating the quality of attention you would give in a face-to-face conversation.
It means being able to make a judgment that says: I know this because you told me; I heard this because you requested that piece of information; you demonstrated interest by joining that webinar; and therefore, based on sound logic, you will get value from what I am about to share.
If you cannot draw a direct line between someone’s signal and the value of your response, the personalisation is lost before it starts.
Trust is the Real Constraint, Not Technology
Underneath the personalisation debate sits a more complicated problem — trust. Buyers have been over-personalised, mis-targeted, and served content that claimed to know them and clearly did not.
The natural question a buyer now asks is: “What would I actually believe anymore?” Brand reputation is built or lost in the gap between what a message claims and what a buyer can verify for themselves.
That impacts what authenticity means, and it has to be reconsidered through the buyer’s eyes rather than the marketer’s intent. A message can be well-researched, well-targeted, and technically accurate, and still come across as intrusive if it arrives before trust has been earned.
The Gen Z Buyer is Coming. Are we, as B2B Leaders, Ready?
Gen Z buyers require additional consideration. They do not want to speak to anyone unless they are truly ready. They have grown up sourcing their own information with confidence; they do not need you to tell them something, and they do not believe you when you do. Instead, they supplement their own research with peer endorsements. Information must be available, and access must be frictionless.
The long-held stat that buyers are 60% of the way through the buying cycle before they want to talk to you has shifted closer to 80%. For Gen Z, it is closer to 90% of the way through the decision-making process before they want to engage a vendor directly. That requires a rethink of what a personalised, relevant next-best action truly looks like.
The boundary between B2B and B2C continues to blur. We are all consumers, and we bring consumer expectations into our professional lives. The old model of funnels and pipelines, pushing leads through to opportunities in a linear fashion, has been the dominant frame for years.
As buying groups grow more complex, the need to understand individual people becomes more pressing than the need to manage the funnel.
Where AI Actually Fits in the Personalisation Stack
AI now sits inside the stack rather than being bolted onto the outside, and where exactly it fits matters more than whether it is present. This is not simply a question of technology and an automated process.
Say you identify a buyer’s early signal of interest and respond to it with everything the platform is capable of — that responsiveness itself can be what tips a buyer off that they are dealing with AI rather than a person. Handled badly, that is the moment trust is lost.
The signal should determine the best next step, and AI is well-placed to make that call quickly and at scale. But the workflow needs to be defined deliberately rather than treated as one continuous, end-to-end process.
Marketing leaders need to decide, workflow by workflow, what good looks like at each stage, where AI can be trusted to act autonomously, and where a person needs to be in the loop before anything reaches a buyer.
Evidence, Guardrails, and the Cost of Getting It Wrong
You can train AI agents to respond based on what they know. The risk sits in what happens when they do not know enough. Without sufficient evidence behind an answer, AI will fill the gap itself, and a fabricated answer reads exactly like a correct one. Nobody spots it until a person flags it, often after it has already reached a buyer.
The fix is not more caution about using AI — it is about training and guardrails around it. A simple rule holds up well in practice: if the system does not have the information, it says nothing rather than inventing something plausible. That single rule protects the brand’s reputation more than almost any other decision in the deployment.
The Dark Funnel Problem Nobody Has Solved
Much of what marketers now call personalisation is built on the dark funnel — the behavioural signals that sit outside any tracked, attributable channel.
The difficulty is that these signals cannot be reliably integrated into a single coherent picture of a person. You can observe, and you can capture, but every inference drawn from a partial signal carries a “what if” attached to it.
Where AI adds real value is not in pretending those “what ifs” do not exist. It is in evidencing and enriching a signal before anyone acts on it, and in being honest about the confidence behind an inference. Anything that can be attached to a signal — a firmographic detail, a previous engagement, a stated preference — helps define a genuine opportunity rather than a guess.
Personalisation on its own is not the goal. The better question is: what is the next-best action for this person, right now, and how does the available signal help define it? Framing the problem that way keeps AI focused on a decision that helps the buyer, rather than on demonstrating how much the organisation knows about them.
Humans still have a defined place in that flow. Where the model’s confidence is low, where the stakes of getting it wrong are high, or where a relationship needs a human tone, that is where a person needs to be involved in the process — not catching mistakes after they have already reached the buyer.
Give People the Choice
Trust rebuilds fastest when people feel they are choosing rather than being profiled. Rather than mandating a response based on inferred intent, offering a genuine choice, framed as “you might also like,” hands the decision back to the buyer and reduces the risk of an overconfident inference landing badly.
This also means deciding, deliberately, whether a programme relies on the dark funnel’s inferred behavioural signal or the white funnel’s information that the buyer has knowingly and willingly given. Both have a place, but they carry different obligations and different margins for error.
None of this can be delivered without addressing the people inside marketing organisations. The skills and capabilities picture is changing rapidly, and marketing leaders have an obligation to help their people navigate that shift rather than be overtaken by it.
The obligation of any marketing leader is to educate and to ensure teams understand the practical applications of the AI tools they now have access to, through use cases, simple how-tos, and shared learnings that make the new capabilities tangible and useful rather than abstract.
That’s the challenge over the coming years: balancing the constantly expanding capabilities AI creates for delivering a hyper-personalised experience with the management and maintenance of trust and reputation. It absolutely can be achieved, but it has to be purposeful and precise. Are you ready? Really?
ALSO READ: Engagement-First Businesses Need More than Just Connected Technology
