I’ve been thinking a lot lately about what it means to publish something and have it matter. Not just be seen — but actually land with the people it’s meant for, and be believed.
We got a small but meaningful reminder of this earlier this year. OTT’s State of the Sector 2025 report has been downloaded over 1,300 times — a number that doesn’t happen by accident. It’s the result of sustained communications work: the right framing, the right channels, following up at the right moments. And it felt good. It still feels good. But sitting in a session at the OTT Conference 2026 about trust in the sector, I found myself asking a question that complicated that satisfaction: how many of those 1,300 downloads came from people who already trusted us? And how do you reach — and convince — the ones who don’t?
Part of what’s made this harder is obvious to anyone working in communications right now. The internet is noisier than it’s ever been, and a meaningful chunk of that noise is AI-generated content that looks authoritative, reads fluently, and is often just wrong. The tools that were supposed to help people find good information have made it significantly harder to know what good information even looks like anymore. For organisations doing serious, rigorous work, that’s a real problem — and it doesn’t get solved by writing better subject lines or posting more consistently.
The number that stayed with me
The session — Your Research Is Good, But Does Anyone Know That? Building a Trust Infrastructure for Think Tanks, convened by Toby Green from Policy Commons and Nick Cleveland-Stout from the Quincy Institute, opened with a figure that surprised me. Only 28% of politically engaged people trust think tanks. Not the general public — people who are already paying attention, already looking for credible analysis. The trust deficit isn’t with disengaged audiences. It’s with exactly the people that think tanks are trying to reach.
A lot of what I do at OTT is help translate complex ideas into content that travels — social media posts, monthly newsletters, blogs. The assumption behind all of it is that the work is good, and the job is to get it in front of the right people in the right format. But that 28% made me wonder whether the problem runs deeper than reach. Whether audiences aren’t just hard to find, but increasingly hard to convince, and whether the sector has done enough to earn their confidence in the first place.
Credibility is built before the publish button
A key takeaway from the session was about peer review — specifically, that think tank reports occupy an awkward middle ground between academic literature and journalism and often get the worst of both. Too informal to be taken as seriously as journal articles, too dense to travel like good journalism. The term that came up was “grey literature” — a category that functions almost as a dismissal, implying that because the review process isn’t standardised or publicly visible, the rigour might not be there either. +
What stood out to me is that this is as much a communications problem as a structural one. If the review process exists but isn’t documented in a way that readers can see, it might as well not exist from a trust perspective. And if comms teams are only brought in once the research is ready to publish, we’ve missed the window to shape how that credibility gets communicated.
Being embedded earlier — asking how a piece was reviewed, who signed off, what the governance looks like — isn’t scope creep. It’s the part of the job that actually builds trust rather than just managing its appearance.
Transparency as a structural commitment
The session also surfaced something that’s been sitting uncomfortably in the background of many sector conversations: funding opacity. Around 40% of top US think tanks disclose no donor information — including organisations that regularly testify before Congress and shape policy. The cumulative effect on public perception is sector-wide. Organisations that operate with genuine independence end up carrying reputational weight they didn’t earn.
The response to this has to go beyond stating values. It has to be visible — published finances, clear governance, documented processes, and communications that make all of it legible to a non-specialist audience. That’s a communications function, and treating it as one changes where and how comms teams need to be involved.
The “Trustmarc” idea — and what it means for AI
The most concrete proposal from the session was a digital trust mark: a clickable, machine-readable token embedded directly in content — reports, blogs, podcasts — that certifies authenticity across four dimensions: accountability, governance, correction mechanisms, and verified identity. A pilot with the Policy Commons platform has already applied these marks to over 373,000 items.
The AI angle is what I found most interesting. At OTT, we’re already seeing this shift in our analytics: ChatGPT and other large language models are increasingly appearing as traffic sources for our website, competing with and often bypassing Google Search or social media. People are going to an LLM first, getting an answer, and either stopping there or arriving at our content through a route we didn’t anticipate and can’t easily influence. Large language models currently have no way to distinguish between a verified institutional report and a confidently written piece of AI-generated content. They treat all web content with roughly equal weight, which is part of why the information environment feels so degraded right now. Trust marks would give AI systems an explicit signal — a way to surface credible, certified content rather than just whatever ranks highest or sounds most fluent.
For think tanks trying to remain relevant in an AI-mediated world, this matters enormously. The infrastructure built around content — how it’s tagged, described, certified — will increasingly determine whether it surfaces at all in the answers people get from the tools they rely on daily.
What I’m taking back
Getting 1,300 people to download a single report is meaningful. But it’s a starting point, not a destination. The harder, slower work is making sure that when those people arrive at OTT’s research — or any think tank’s research — they have genuine reasons to trust what they’re reading, reasons that are visible and verifiable rather than just implied by reputation.
That work starts earlier in the process than most comms teams are currently involved, and it requires asking questions that sit closer to governance and research design than to copywriting. Getting into those conversations feels like the most useful thing I can do right now — and the most honest version of what communications is actually for.