For campaigns: before, during, and after. Before, to set strategy and a baseline: find the audiences that matter and the organic roads in, so spend is more effective and efficient. During, to see what's genuinely moving and adjust while you still can. After, to measure precisely the impact of your effort, with recommendations for the future.

For reputation: ongoing, focused on the conversations you care about — a brand narrative, a policy fight, an emerging risk — so you can proactively build credibility and prevent crises from escalating unnecessarily.

And as a post-mortem when something has already happened: what happened, what may have happened differently, and what difference your response made.

An added bonus: the platform becomes more tailored to you with every analysis.

Causal effect: how much of a conversation, a narrative's spread, or a shift in sentiment would not have happened without a given account, message, or campaign. We use this information to help companies make better decisions and understand the true impact of their campaigns. 

Social listening tells you what's being said, how often, and by whom — usually ranked by volume, which means the loudest accounts dominate every report whether or not they changed anyone's mind. We answer a different question: not “what happened near your message?” but “what did your message cause?” That requires modelling the network and the counterfactual, not counting mentions. The two are complementary — listening describes the conversation; we attribute it.

Narrative mapping, sentiment analysis, behavioural analysis, and narrative dashboards describe the conversation. Some of them are repackaged social listening; the better ones add structure and sentiment. All of them stop at correlation: things that moved together, with no way to say what caused what. And a correlational tool cannot tell you what to do, because action needs causes (and causality is the only basis for action). We measure and add context around true influence, so navigating online conversations stops being guesswork.

Yes. That's what the mathematics shows: causal influence reflects an account's position in the network and the trust it carries, and that persist in future observational windows. Our validation studies confirm it — accounts with high measured influence predictably go on to cause future conversations.  

No. The underlying technology is not generative AI. It's frontier mathematics: patented causal algorithms that compute influence on social networks, a problem long thought mathematically impossible to solve. Our underlying technology was developed by mathematicians and data scientists at some of the world's top research institutions, rooted in over $15m of research, in the same causal-inference tradition recognised by a recent Nobel prize in economics. It's statistical modelling with explicit, inspectable assumptions. Gen-AI help interpret findings, with a human in the process, and we indicate wherever it's been used. 

Publicly available social media activity: public posts, public interactions, and the public connections between accounts. We do not access private accounts, direct messages, or restricted content.

Yes — and, more usefully, we tell you whether these campaigns are effective. Coordinated amplification is engineered to inflate exactly the metrics most tools rely on, which is why those tools struggle with it. Causal structure is harder to fake: activity that doesn't propagate along real relationships, or that appears simultaneously without shared connections, doesn't behave like genuine spread. But detection is only half the question. Plenty of coordinated activity is loud and causally inert — it circulates inside its own network and moves no one. We measure the campaign's actual effect on the conversations you care about, so you can distinguish an operation that needs a response from one that isn't going anywhere.