graph colored graphic

Easily set up the most granular social monitoring in the world

Our technology reads social media & news feeds, extracts the core opinions and claims you need to care about.

Most media monitoring companies focus on sending you alerts about a topic you set up a keyword search for along with some basic sentiment analysis. They look a bit like this →

british airways graphs

Stop drowning in a mass of mentions

Companies rely on focus groups, data scientists, and market analysts to trawl through content and identify key industry trends and product ideas. These take too long to find insights, and involve reports which are often unreliable, slow, expensive, and inefficient.

The majority of social listening tools only alert you when your brand is mentioned, without letting you delve into the “why”. They often only work on one type of data source e.g. Twitter, and offer very basic analysis like word clouds.

Producing actionable insights is still a manual task, meaning hours of scrolling and the risk of human bias in the analysis.

Don’t let your team fall into that trap.

We use Natural Language Understanding to extract relevant opinions about any topic using 3 simple steps:


Organise material by sentiment and themes across multiple types of (open) content


Clustering similar arguments to track and assess ongoing narratives


Online Opinons visualizesed for strategic insight and action

screenshot photo of
opinions count line
flu vaccine awareness

Our AI can analyse millions of data points across multiple channels in seconds

Let our AI do the hard work analysing the data so you can spend more time on insights and strategy. Here’s the benefits of Factmata’s Narrative Monitoring technology:


Save hours, or days, infinitely scrolling through the mass of mentions to understand what is being said about a brand or topic in seconds.

Early Warning

Our AI can identify narratives as small as a three similar opinions, even in topics containing hundreds of thousands of opinions, meaning you get early warning of potentially harmful narratives or opportunities.


Our AI doesn’t have the same biases as a human to political content or to patterns which might skew summaries or filtering conducted by humans on the same data.


No more spreadsheets, you can get all the graphs you need to turn our qualitative insights into quantitative evidence of ROI.

Cluster similar opinions into narratives

There are many ways to express the same idea, and different people might be saying the same thing in different ways and using different figurative language. We combine similar statements into clusters of ideas that grow over time, from the moment of detection.

opinions count line

Rank narratives on multiple dimensions of risk

In the final step, we are able to analyse and rank narratives in unique ways, according to their linguistic style.

Propaganda score

Likelihood of a narrative being politically biased


Level of deceptiveness in the way a narrative is phrased


Number of social interactions (likes, shares, retweets) for a piece of content

Threat score

Likelihood of a biased narrative being widely spread. The metric takes into account the propaganda score and popularity

Track a narrative from when one person starts it, and pinpoint its origin

Each narrative can be dissected into its component opinions. With Factmata, you can see who made them, where they spread to, the individual stance, and characteristics. Track when and where they started, how they were expressed, and how they grew over time.

price is too
this isnt vegan
jonah goldberg social

See who is involved in each viewpoint

Quickly find key opinion makers of every stance and narrative. See who different messages resonate with and who is the source of interesting opinions.

A unique AI for personalised needs - train your personal AI to improve every day

Factmata can be adjusted to suit the specific situation. Every feed is unique because you as an analyst help to train it. Sometimes narratives we find were not very threatening or relevant, had the wrong name, were incorrectly clustered and had irrelevant claims within them not matching the narrative, and so on. You can edit any of these things.

is narrative interesting
social media shares

Track a claim as it spreads across one platform to another

Our engine is able to track narratives from multiple sources

We can ingest data sources from any provider, including:

We can also provide data via our own partners from sources including:

Our technology has a number of core advantages.

That’s because it’s been cultivated by a team of dedicated engineers, and is driven by a community of experts. Factmata is the only known system that can differentiate between texts expressing harmful views, and texts reporting on those views. This is built upon our proprietary quote detection model to distinguish direct vs reported speech.

Factmata does not replace people. Experts make our technology better. People use our tools to make better decisions.

True Language Analysis. We can understand content for its potential offensiveness, incorrect information, controversiality, editorial quality, slant and racy opinions, using experts in assessing language, opinions, news stories and more.

Combining domain experts and AI. We use our own communities of journalists and subject matter experts for niche topics to help train and enhance extremely detailed classification algorithms on your content.

"So many listening tools still leave you asking "so what?" Factmata doesn't leave you hanging, it gets straight to the heart of who and what is driving the conversation."

Alex Judd, Head of Impact and Planning at Clarity PR

Our Natural Language Understanding system is always learning

Internet Content

We monitor content from a vast network of sources, and update our database of harmful content daily. That means our AI learns from a comprehensive base of content, which is always up to date with the latest slang and rhetoric.

Communities of experts

While AI is often trained on data labelled by experts in a one-off process, our experts continually refresh training data and give feedback on the content the system finds, to make sure the algorithms can keep up with the ever-shifting political and media landscape.

Internet users

We let anyone give feedback on the AI’s output, which ensures our AI remains relevant and free of bias itself. We have strict controls and can monitor when we are being attacked or trolled to game our classifiers. If you want to be shown instead of told, you can test out our API now.

Our AI powers Content Score, which processes millions of pages of internet content every week for continuous learning

Our technology has won multiple awards

ircai international research

2021 Global Top 100 Company

msduk uppercase black

2021 Advanced Digital Solutions Winner

globee awards seal

2021 Globee Best AI product in marketing winner