Find out who is losing trust in your brand

The task

A trust score that’s falling tells you something is wrong, but it doesn’t tell you where exactly your problem lies. The same three-point drop can be a broad erosion across every audience or a sharp pullback in one group that happens to be large enough to move the average, and those two situations call for different responses: one is a messaging problem, the other is a targeting problem.

Morning Consult’s AI Connector can address this problem in one conversation. We track trust in thousands of brands daily, across the full set of demographic, political and economic cuts, which enables our clients to keep a close eye on their brand health. Instead of exporting a dozen crosstabs and comparing them by hand, the AI Connector lets you ask who is losing trust. You get an answer in the form of a profile: which groups moved, by how much, and how they compare with the audiences that held steady. For a communications or reputation team, this helps you go from knowing that trust fell to knowing who to talk to about it.

Prompt

Using Morning Consult data, can you profile the consumer losing trust in [BRAND]?

The result

The AI Connector pulls Morning Consult’s trust tracking for the brand and returns a portrait of the low-trust consumer. For X, the profile in the third quarter of 2026 was sharply defined. Net trust sat at minus 11.6 among women versus plus 4.8 among men. It split hard by party: plus 20.1 among Republicans, minus 11.5 among independents, minus 18.8 among Democrats. It was negative at both ends of the age range (minus 10.9 among 18-to-29-year-olds, minus 15.2 among those 65 and older) and roughly flat in between. And it was negative or flat in every income band under $150,000. Each cut came with the weighted interview count behind it and a chart, so the profile could go straight into a reputation briefing without a second pass through a spreadsheet.

Narrowing the question

The more useful question for a response plan is usually who is losing trust fastest, which is a different group from who distrusts the brand most. A follow-up prompt makes the distinction explicit.

Prompt

I’m less interested in trust overall and more interested in who is losing trust the most.

With the AI Connector, you can reframe the analysis from current level to change over time, ranking audiences by their decline from their peak to the current reading. For X, the steepest declines came from the groups that trusted X most a year earlier: households earning $100,000 to $150,000 fell 18 points from their third-quarter 2025 peak (plus 13.3 to minus 4.9), $75,000-to-$100,000 households fell 15 (plus 15.3 to plus 0.5), 30-to-44-year-olds fell 13 (plus 15.1 to plus 2.6), and men fell about 12 from their late-2025 high of plus 17 to plus 4.8. Democrats, who had recovered to minus 6.3 in the third quarter of 2025, slid back to minus 18.8. Meanwhile the groups with the lowest trust (women, those 65 and older, lower-income households) had moved far less, because they had little trust to lose, and Republicans held steady at around plus 20 throughout.

That is the actionable output: it separates audiences where trust is low but stable (a long-term positioning issue) from audiences where trust is falling now (the group a campaign or statement should address first). For X, those are two different sets of people. From here, the natural next questions are:

• “Chart the trust trend for the three fastest-declining groups over the past year, on one chart.”

• “Does the same pattern show up for [COMPETITOR]? I want to know if this is us or the category.”

• “Write a one-page summary for our communications lead: who is losing trust, since when, and what we don’t yet know.”

Keep in mind

Ask for change, not just level. Most trust questions get answered as a snapshot, and a snapshot can’t tell you where to act. The follow-up that reframed “who distrusts us” as “who is losing trust fastest” is the single most useful move in this whole exercise, and it applies to any metric: favorability, consideration, purchase intent. When a number is moving, ask who moved it.

Make sure to check the sample sizes. A dramatic finding in a small group is often just a small group. Ask “how many interviews is that built on?” before leaning too hard on a figure’s finding.

Put the competitor in the same question. Trust in X can fall because of X, or because of what’s happening to social media. Adding “and show the same cuts for [competitor]” to your prompt turns a brand finding into a category finding, which is usually the version your leadership team wants.