A product manager at a mid-market SaaS company told us her team's new AI-personalized dashboard was technically a success. Engagement on the personalized modules was up. Support tickets were also up, from users asking some version of the same question: "how does it know that, and can I turn it off?" The feature was accurate. It was also opaque, and opacity is what turned a relevance win into a trust problem. The same pattern shows up in a fintech app that quietly reorders investment options based on inferred income, a biotech patient portal that surfaces different content based on a diagnosis code nobody explained it was reading, or a media site that reshuffles a homepage based on browsing history the reader never agreed to have used that way. None of these products did anything unusual. AI-driven personalization is now table stakes. What separates the ones users trust from the ones they quietly resent is not the personalization itself. It is whether the product tells you what it is doing with your data while it does it.
Teams have spent years optimizing personalization for relevance: better recommendations, better targeting, better click-through. That work is largely solved at the model level. The open problem is different: whether users trust the product enough to let it keep personalizing. Cisco's 2026 Data and Privacy Benchmark Study, which surveyed organizations on privacy and AI, found that 87 percent of respondents agree that strong privacy protections make customers more comfortable engaging with AI applications, and that transparency, specifically giving people clear information about how their data is collected and used, is the single most effective lever for building that comfort, ranked well above compliance messaging or breach-avoidance language, according to Cisco's own benchmark study. In other words, the thing that makes people comfortable with AI personalization is not the accuracy of the personalization. It is whether they were told, plainly, what is happening.
This is the same shape of problem we described in our piece on AI explainability, but it is a distinct one. Explainability is about helping a user evaluate whether a specific AI output is correct. Personalization trust is about whether a user understands and consents to what data is feeding the experience in the first place, before any single output even appears. A recommendation engine can be fully accurate and still lose trust if the user has no idea their location, past purchases, or inferred health status are shaping what they see.
Two B2B SaaS companies in adjacent categories both shipped AI-personalized onboarding last year, surfacing different setup steps based on a new user's role, company size, and early clicks. The first company personalized silently: users simply saw a different flow than their colleagues, with no explanation, and support fielded a steady trickle of "why don't I have the same options as my teammate" tickets. The second company added one small element: a dismissible line near each personalized recommendation reading "Suggested based on your role and the integrations you connected," with a link to see and adjust what data drove it. Adoption of the personalized flow was statistically similar between the two products. Trust was not: the second company's post-onboarding survey showed meaningfully fewer users describing the product as "watching me" in open-text responses, and its support volume on personalization-related confusion dropped by more than half. Same underlying model, same accuracy, one visible sentence of difference.
This is Aero's own working framework for auditing a personalized AI feature before it ships, not an established industry standard. Five questions we ask a product team:
A team that answers "no" to two or more of these is likely optimizing personalization for relevance while quietly spending down the trust it needs to keep using that data at all.
Personalization transparency is closely related to, but distinct from, the regulatory disclosure work many teams already have underway. We covered what the EU AI Act's Article 50 requires when AI-generated content itself needs labeling in designing AI transparency and disclosure. That obligation is about disclosing that AI produced something. Personalization trust is about disclosing what data an AI system used to decide what a specific user sees at all, a design problem that exists whether or not the output itself is AI-generated, and one that shows up just as often in a recommendation feed or a reordered pricing page as in a chatbot response. Teams that treat these as the same checkbox usually satisfy neither one well.
No. A disclaimer buried in a privacy policy does not change how a user experiences the product in the moment. What moves trust is a short, specific, in-context explanation next to the personalized result itself, plus an easy way to adjust it.
Some will, and that is the point: informed consent is worth more than silent compliance. In practice, most users who understand and can control personalization keep it on. The ones who would have turned it off anyway were never going to trust an opaque version either.
Run the five-question checklist above against your single most heavily personalized surface, the homepage, dashboard, or recommendation module your team is proudest of. If you cannot answer the first two questions in one sentence each, that is the feature most exposed as user expectations around data transparency keep rising.
Curious whether your AI personalization is building trust or spending it down? Talk to Aero.
AI personalization is now standard across SaaS, fintech, and healthcare products, but accuracy alone doesn't earn user trust. Here is Aero's practical framework for designing personalization that discloses what it's doing instead of quietly guessing.
Most AI features are judged on accuracy, but users decide whether to trust an output long before they can verify it is correct. Here is Aero's practical framework for designing AI explainability, and why that is a different problem than the disclosure rules already on the books.
AI copilots and agents are becoming standard issue across SaaS, fintech, and biotech alike. Here is Aero's framework for what still makes a product distinct once every competitor has shipped the same feature.
