Orchestrated, earned, audited: A vision for B2B trust in 2026, from Spotlight Summit

A historic hotel in downtown Kansas City isn’t exactly where you might expect the future of trust in B2B to be defined, but that’s arguably what happened last week at Spotlight’s Summit event, where influence, advocacy, authority and the role of LLMs in defining this new era were explored from every angle.

Picture a B2B buyer narrowing down a shortlist. They ask ChatGPT or Gemini what it knows about a vendor, get an answer, and then do something the machine didn’t ask them to: they check that answer against a Reddit thread, a G2 review, or a phone call with someone who’s actually used the product.

That single habit, ask the machine, then check with a human, turned out to be the idea running through nearly every session at Spotlight Summit last week. Spotlight @is a fourteen-year-old B2B research and analyst relations company, and the report behind the event, Influence Orchestration, draws on 23 interviews with leaders who build and measure B2B influence, 27 published studies of buyer behaviour, and an analysis of millions of AI engine answers and citations.

I sat down with John Rockhold, Spotlight’s VP of Market Engagement, immediately after the event for a recent sponsored episode of the Trust & Influence in B2B podcast, to unpack what the report found. The loop he described doesn’t fire once, at the start of a buying process. It runs continuously, through the whole evaluation, as buyers work out who’s actually worth trusting.

A loop not a gate

That loop is worth sitting with. Rockhold’s research shows buyers checking machine-generated answers against vendor sites, search results, and each other, repeatedly, throughout the process of narrowing down who they’ll actually talk to. “The robots really prefer humans to answer questions for humans,” Rockhold told me, and it’s a finding with a genuine sting for anyone who assumed AI answer engines would simply reward polished, well-optimised vendor content.

They don’t, not on their own. What they reward is content that traces back to a real person saying something they actually believe, and when buyers aren’t satisfied with what the machine has told them, they go further still. “They phone the analyst, they talk to an influencer, they talk to a peer at a different company that has a similar conundrum, they talk to fellow users within the same company,” Rockhold said.

“To seal the deal. To feel better.” That loop also runs at both ends of a decision, not just the middle. Anjali Yakkundi Yakkundi, SVP of marketing at Movable Ink and a former Forrester analyst, put it neatly in Spotlight’s report: analysts now get consulted “more of the discovery, ‘Oh, I forgot this vendor,’ or at the very end: ‘Did I forget anybody?'” Early or late, the mechanism is the same. Something a human said gets checked, and something a machine said gets checked against a human.

Evaluation and validation aren’t the same job

That distinction, between what a machine will cite and what a human will trust, was the subject of one of the sharpest exchanges at the Summit. Bijou B. (Barry), AI principal analyst at G2, framed it this way in Spotlight’s report: “Owned versus earned is about evaluation versus validation. And those are two very different jobs. An evaluation gives me the details of what I’m measuring as a form of trust. A real review gives me context: what I expected it to do, and how it actually performed. One doesn’t become more important. But one cannot happen without the other.”

It’s a useful correction to a debate that too often gets flattened into owned content versus earned content, as though marketers have to pick a side. They don’t. They have to understand which job each one is doing, and stop asking owned content to do earned content’s work. It also explains a pattern in the report’s own data: buyers who get a vendor recommendation from an AI tool still return to that vendor’s own website close to half the time, sometimes considerably more, to check the specifics against what they were told. Owned content isn’t dead. It’s doing a narrower, more specific job than it used to.

Earned beats owned in B2B trust

That’s where the report’s more provocative claim comes in: earned beats owned, at least for what the machines weigh most heavily when deciding who to recommend. I’ll admit a personal stake in this one. The report quotes something I said in an earlier interview with Rockhold’s team: “What separates competent from good is the authentic voice of somebody you trust, and that’s an earned conversation.” Rockhold explained why that line closed out the report’s section on owned content.

“We’re trying to be very careful to not make it a black and white scenario,” he said. “Owned content still has its place, still has its merits, still has great value. But what people say, what third parties say, is what truly takes your visibility to the stratosphere.” His practical advice is simpler than it sounds: point your owned content at what other people are already saying about you, and you’ve turned owned content into earned content without losing either. It’s not a wholesale rejection of the last two decades of content marketing so much as a correction to what that content was ever capable of doing on its own.

The ceiling on owned content

The data backs the direction of travel, if not a single precise tipping point. Spotlight tracked 915 vendors across ten B2B categories and six AI assistants, and found that visibility collapses once a vendor’s owned content passes roughly half of what backs them. Below that line, some vendors still reach 50 to 70% visibility.

Above it, the best anyone manages falls to the floor. It’s not a reason to abandon owned content. It’s a reason to stop treating it as the whole job, and to start treating the ratio between what you say about yourself and what others say about you as something worth actually managing.

Nobody owns B2B trust, yet

The harder problem, and the one I think matters most for anyone running a B2B marketing function, is that almost nobody owns this internally. Rockhold’s team asked every interviewee, unprompted, what influence meant to them, and trust came up nearly every time. “You can’t have influence without trust,” he said. “They go hand in hand.”

But responsibility for building that trust is scattered across analyst relations, communications, customer references, customer engagement, and influencer relations, often with no one accountable for whether those functions are saying consistent things. Machines notice inconsistency. Buyers notice it too.

Rockhold’s report goes as far as sketching a “VP, Influence” role, complete with a salary band, as one plausible answer, alongside two other predictions: that vendors gaming the machines are already on borrowed time as the AI platforms clamp down on inauthentic content, and that by 2027 industry analysts will increasingly ask for customer proof rather than taking a vendor’s word for it, the way Gartner already turned off vendor-supplied references in favour of its own peer review data.

Whether the exact title, or the exact year, lands inside organisations on the report’s stated timeline is a reasonable thing to be sceptical about. That something has to own this, rather than everything owning a slice of it, is harder to argue with.

What struck me most, talking to Rockhold after two days immersed in AI answer engines and citation data, wasn’t a claim about the technology at all. Asked what he’d taken from the Summit, he came back to something almost old-fashioned: how much people still valued talking to each other in person. “It should be human driven and AI in the loop,” he said, reversing the phrase most of the industry has settled on. “AI riding sidecar. Human driving the bus.” It’s a fitting way to close the loop this piece opened with. The machine asks and answers, but the trust still gets settled between people.

Three things to check this week to improve B2B trust

• Ask two or three different large language models what they know about your company and your category, and look past whether you’re mentioned to what’s actually backing the answer. Check the sentiment of it, the veracity of it, and how recent it is, and if the model doesn’t show its sources, ask it to. If a model is only pointing back to your own website, that’s a gap worth taking seriously in its own right, not a reason to simply publish more of the same content faster. If what you find is good, that’s not a reason to relax either. It’s a head start worth building on before someone else catches up.

• Look honestly at your owned-to-earned mix, not just whether you have both, but roughly what share of your visibility either one is actually responsible for. Rockhold’s report points to a rough industry default of about 75% owned, 25% earned, and argues that ratio needs to flip, not disappear. If the bulk of what represents you online is content you wrote yourself, you’re closer to the point where Spotlight’s research shows visibility collapsing than you might think, and the fix isn’t more owned content, it’s more of other people saying true things about you.

• Name who actually owns influence inside your organisation today. It doesn’t need to be a single department doing all the work, but somebody needs to be accountable for whether analyst relations, communications, customer references, and influencer relations are pulling in the same direction. If the honest answer is nobody, or everybody a little, that’s not a communications problem to solve later. It’s the first thing worth fixing, because the machines are already treating your inconsistency as a signal, whether anyone inside the business is watching for it or not. And if what you find is a mess, that’s not a reason to panic. It’s a reason to start.

Links below to listen to the podcast episode in full and to download the report.

https://spotlight-io.com/insight/2026-influence-orchestration-report