Atlas Solutions

High impact content, powerful advise

Financial Stories

Nothing like the financial media could be a better fit for our products. Our solution comes as a natural fit to this landscape. There are links between entities which are in plain sight but are often ignored. Links which have the power to distinguish between the noise and the story. Whether these links are of a financial nature or a political one, Atlas 4 gives our clients in the financial media the ability to see them and incorporate it in their content.

The financial media can use our solutions for two purposes: To create contents or to generate financial advice for their readers. Our data is designed to be rich in context and so it can provide leads and ideas not easily discovered elsewhere for both purposes . What's more is that our AI technology behind all this is highly explainable. That's why any idea or advise coming out of our products can easily be used in the publications.

Our product is very felxible when it comes to usage. If they choose to, our clients can use an AI-assisted visual interface to find story leads or advise ideas. Alternatively, our intelligent metrics can be used in order to analyze financial assets with more depth and context to provide readers with high impact stories and reliable investment advise. Let's take a closer look at each of these options.

Asset 360°

One of the easiest ways to navigate through financial assets is by using our intelligent visual interface which we call Asset 360°. This interface puts on display the context of a financial asset. The user can find what other entities are related to an asset. Each of these direct relations have their own, separate relations to other entities. These are also visible to the user through interaction. By clicking through this interface, a journalist can uncover some long reaching and overlooked connections between entities (individuals, corporations, political movements etc) and form content or advise related ideas. For example, the user may start with an asset of interest and end up discovering the political inclinations of its leaders and link it to an upcoming election. As another example, assuming that the asset under investigation is a company, the user may end up discovering that a product of this company is used by another company which they did not expect.
When navigating through the interface, the user can check charts, graphs and AI metrics which are positioned in a semantically relevant manner to the research topic. This boosts their reasoning power and helps the user to provide solid support for their choice of content.

Intelligent Metrics

One of the issues data-driven journalism is faced with is the abundance of data. There is simply too much data out there and it is very hard to find the untold story from it. Metrics can significantly simplify this by giving the journalists the ability to see which entities matter the most in their case.

One of the metrics which could help with this matter is a simple concept at its core: semantic similarity. We call it semantic because it measures the similarity very much like humans do. This metric compares all the entities considering the data they generate, their explicit or implicit relationships to other entities and the data generated by these relted entities. In this way, the metric captures the context of each entity by using all the knowledge there is about them. One interesting aspect of this metric is that it is not restricted to entities of the same nature. For instance, it provides the similarity between a company and a government, an individual and an event and so on. This gives rise to powerful analogies which are great to have during story generation. Apart from that, this metric also enables data-driven journalists to cluster entities using objectives they have in mind. Like clustering companies across political spectrum.

As another example for metrics, we calculate a propagation score for each entity. Let's say an event has happened and we want to quantify how far the effects of this event will go and which other entities it will touch. Using traditional approaches, coming up with this figure is hard and time consuming (if possible at all). This is while using automated reasoning, this is organically computed. This is because our knowledge graph knows which entites are connected to one another, what the strength of the connection is and how distant one entity is from the event for which all this computation is done. Using that knowledge, our Semantic AI algorithm will traverse this network and calculate the impact of the event for each entity it encounters. In other words, the algorithm reasons with the knowledge graph in a similar way to human reasoning, only at scale.

Inversely, sometime we want to know the exposure of an entity to another entity. For example, we want to know how much a certain country is exposed to climate events and in which way (positive or negative). Again, similar to human reasoning, the AI algorithm will go into a reasoning process. While the country we're tracking may get a direct hit from the climate event, it can also be indirectly impacted by the event as some of its dependencies can go under the influence of this climate event. What's more is that these separate impact quantities can have synergy between them resulting in a combined effect larger than the sum of individual impacts. If so, it won't go unnoticed by the AI. To do all this manually is extremely time consuming and the results will vary from one group of analysts to another. To do this using traditional approaches is also time consuming as the data linkage phase will take a lot of time (months in most cases).


The list of metrics which could help data-driven financial journalism doesn't end with exposure and propagation. We generate metrics like Social, Political and Environmental scores for each asset to keep track of the various ways they impact the economy. Or Synergy which detects if the union of two entites (like two companies) will have an impact higher than the sum of the impact of each entity. Core metrics like Centrality helps data-driven journalists to trace developments to their source(s).

Where Are We Heading?

In addition to all this, we also provide the time-series values of these metrics. But this is not your average time-series. We say that, because embedded in it, you can see the trajectories of each entity with respect to other entities. For example, it will help you to see whether a company is leaning from the left to the right of the political spectrum over time. As another example, it enables you to track and see whether two resource rich countries are converging into an armed conflict over time which could impact supply chains for several assets. You can do that by simply comparing the semantic similarity between the entity and armed conflict (which is another entity) and see if the similarity increases over time. The same logic can be applied to all the metrics in our inventory. You can check to see if an entity's exposure is increasing or declining and so on.

If this picks your interest, contact us to arrange for a demo or in depth discussion about your stories.

Generic placeholder image

Founder & CEO

March 1, 2021