Module: InterRed ContentAgents

AI & intelligent Knowledge Management

The amount of information that we have to deal with on a daily basis has risen constantly for some years now. But information itself is initially useless. Only information that is linked to a meaningful context can become knowledge, and thus one of the most important competitive factors of our age.

Knowledge is information that is useful in a given context. Examples are the CV of a person named in an article, other products that "match" a product, or recommendations based on a user's behaviours. Many companies have huge stores of information that lie dormant and unused in their IT systems. With intelligent software such as InterRed ContentAgents, this data can be transformed into thematically coherent knowledge that can be exploited.

InterRed ContentAgents as an AI component are both an innovative recommendation system and a smart search tool.

It first analyses the basis of the information. This can consist of the contents of the InterRed multi-channel publishing system, but can also include, for example, documents from the database or the content of “foreign” sources of information. The analysis is completely autonomous based on the predefined sources. InterRed ContentAgents identifies the topics using the contexts of the various texts. It autonomously learns to distinguish between issues related to the court (as in the royal family), courts (as in the justice system) and courts (as in places where certain sports are played). So, in our example, items related to the Queen would be assigned to "the court" (royal family), but not to the courts of justice or courts used in sport.

The result is a content-based, semantic knowledge web that is further refined and adjusted as new content is added. This web then provides the basis for a wide range of ContentAgents applications. The fact that it learns on its own, constantly adjust its knowledge webs and automatically pick up and integrates new topics makes ContentAgents clearly superior to manual, hierarchical systems such as ontologies, categorisations or keyword lists. In day-to-day practice, one quickly realizes that its automation delivers significantly better results than classification systems that are manually maintained and updated. The lack of manual effort very quickly amortises the cost of Content Agents.

What benefits does InterRed ContentAgents offer InterRed users?

InterRed ContentAgents learns, among other things, the relationships between texts and associated concepts. The benefit is extremely versatile allowing us to constantly discover new fields which we research and develop new agents for. The system is thus able, for example, to identify on its own which texts are related and offer them to the user. This is of tremendous help in all relevant editorial processes, for example, in planning the topics to be covered on the day, content research related to a new product, or simply writing an article. InterRed ContentAgents also recognise people and places and can draw attention to other relevant persons or places in their environment. It can also automatically recognise products and existing information from sources as varied as the news to operating manuals or related products, just to name just a few examples.

And what benefits does InterRed ContentAgents offer end users?

End users often reach our clients' online offers via search engines or visit their websites looking for individual topics relevant to them. They can interact with the day's events in a particular section, but can also conduct subject-specific research, product research, etc. In all cases, InterRed Content Agents automatically recognises the end user's interests because of the search terms they use or the information they read and use this information to offer further, relevant information. This then frequently extends the user's stay on the site and their click-through rate, thus adding significant attention to our clients' portfolios. Our client is perceived as individually significant, paving the way to a business transaction with the end user.

These are just a few examples of the many possibilities offered by InterRed ContentAgents technology. The system is based on the latest information technology. Innovative approaches such as genetic algorithms, fuzzy logic and SNLP turn bits and bytes into intelligent behaviours.

Future-proof Knowledge Management

The InterRed Content Agents offer future-proof knowledge management - both for the user of InterRed as well as for end users.

Specialised agents have been created to provide optimal support for different tasks. The following ContentAgents are currently available.


Modern knowledge management with 'semantic web' technologies. The semantic, intelligent linking of contents. The fully automatic grouping of themes. The continuous, independent analysis of existing contents with the incorporation of new findings.


The ContextAgent is the semantic recommendation system of the ContentAgents. Using the latest methods (Text Data Mining, Concept Detection) the ContextAgent analyses contents and independently finds the 'used texts'. The manual entry of further information (metadata) is not necessary; the ContextAgent works autonomously.

In a similar way to the human brain, the neuronal network of the ContentAgents applies newly gained knowledge (new texts) to information that is already available.

The intelligent, content-based linking of texts enables innovative, intuitive information retrieval as well as an optimum presentation of the existing texts and information. Structuring of large amounts of data (Big Data) enables the recommendation of situation-relevant content in each case.

The automatically generated recommendations increase the information content and the interest of the user and therefore increase the click rate and depth on the specific website. When used in an intranet, the clever combination of old knowledge and new questions make it possible to create something that really is new.


The automatic keywording of texts.


The KeywordAgent analyses contents and independently provides the corresponding keywords. On the basis of the ContentAgents technology it scours the available contents and focuses the selection on significant keywords. This makes manual keywording a thing of the past.


Dynamic updates to innovation and change around our clients' interests.


Identifying trends and key issues early is a major challenge for publishers and corporate communications departments alike. What are my customers interested in? What new topics are developing? Have we already covered these issues? If so, to what extent and under what aspects? What are the others writing about?

Information becomes obsolete much more quickly nowadays. That is why it is even more important to be among the first to recognise the latest topics and cover them intelligently. InterRed TopicMeter is based on the knowledge management technology of InterRed ContentAgents. The TopicAgent observes and evaluates the websites, Twitter feeds and internal sources you tell it to follow. Dynamic updates to topics of interest for our clients can be organised into categories to provide a quick overview through the jungle of information: what topics are currently on the rise, what are today's main keywords, and, above else, who is covering a certain topic how and where? In addition, TopicAgent checks your own, already existing content and articles to help you present your opinion on issues.


The ideal linking of AdServer applications.


The AdServerAgent autonomously recognises subject priorities in articles. It can compare them with relevant keyword spaces for Adserver and deliver relevant keyword spaces for the article to AdServer. Adserver then delivers relevant advertising for the context of the article. If the subject areas change radically, the classifications of articles on the subject are adapted automatically. Manual intervention is not necessary. The AdServerAgent allows constantly updated and targeted management of advertising. The prerequisite for use is manually defined keyword spaces that are coordinated with Adserver.


The recognition of persons.


Irrespective of whether it is politics, business, science or history - it is people who influence the world. The PeopleAgent autonomously recognises the stated persons in a text. This can subsequently be used for a wide range of applications, e.g. for creating glossaries and person-related theme pages and/or for determining the carriers of knowledge in a company. The PeopleAgent puts the person in the centre.


Automatic product recognition in continuous text. The perfect, automated linking of E-commerce applications.


Test and consumer magazines inform and test products to advise buyers with their purchasing decision. The ProductAgent enables fully automated product recognition in continuous text and offers the possibility of connecting these products with external applications such as online shops or price search engines. For the user this results in the possibility of the direct acquisition of the required product. For the operator, the ProductAgent creates the possibility for the commercial sale of either tested or advertised products.

The laborious and error-prone linking of textual contents with product offers can be set up to occur on a fully or partially automated basis.

Requirements: Product database of the customer with information on manufacturers and product designations


The new generation of search engine.


The SearchAgent is the search engine for the ContentAgents. It visualises the results of the ContextAgent, KeywordAgent, PeopleAgent and ProductAgent. The SearchAgent therefore enables optimum information retrieval and offers considerable benefits compared with conventional search engines.

It not only shows the texts that contain the search words which are being sought, it also contains the texts that are related semantically and in terms of their contents, meaning those which may not contain the actual search terms. The SearchAgent also provides the possibility of entering full texts as a search criteria instead of individual search terms in the normal way. Using a Widget which is integrated directly in InterRed, the SearchAgent shows the suitable keywords for every text as well as persons who are named in the text.

In addition to the list display the SearchAgent also has a special screen which is called a 'RelationBrowser'. This presents content-based relationships in a semantic network.

Requirements: ContextAgent, KeywordAgent, PeopleAgent


Building vocabulary with words that are related in terms of their meaning.


The SynonymAgent provides words which are related to another word or a phrase and enables simple and targeted search functions. In this way it supports the user in the search for synonyms like a kind of thesaurus.

The SynonymAgent is optimally suited to 'Semantic Query Expansion' meaning the synonymous enhancement of search enquiries which provide the user with better and more precise results than conventional procedures do.


Associations are terms that are related in terms of their content but which are not, however, synonyms.


In this context 'building' would be a synonym for 'house'. The AssociationAgent analyses the relationships between terms and shows those that are related in terms of their content. In this example, further associations with the term of 'house' would be, for instance, 'property', 'buildings insurance', 'driveway' or 'garden'.

The AssociationAgent therefore enriches the search spectrum by offering its own 'word clouds' and making them usable.


Automatic location identification in continuous text. Identification of local topics.


GeoAgent independently recognises all towns in Germany appearing in text, can also specify the post code of the correct location, and has a radius search feature.

GeoAgent thereby enables the automatic localisation of texts. Benefits are extremely diverse: local topics can be automatically determined and visualised on web portals without manual effort. In the increasingly locally-orientated world of mobile devices, information geographically local to the user's own current environment can be automatically offered in addition to automated thematic content. The functions also provide a decisive advantage in editorial production processes such as research, selection of topics and their linkage, for example in local editorial offices (local newspapers). Like all ContentAgents, GeoAgent also automtically determines locations and their geo-coordinates and links them to a virtual map, so that thematic proximity searches can be implemented with minimal effort.

We are here for you