Identrics. Customisable AI solutions that help you understand data.
Expand your offering with artificial intelligence.
Our ready-made, custom machine learning models help media intelligence providers save time and money, improve their products and expand their tendering. Free up your staff so they can focus on human intelligence-intensive tasks. To discuss implementation of AI and Machine learning in the publishing sector:
Customised models to provide more profound insights.
Our best-of-breed text mining technologies make sense of any type and volume of text in 11 languages. We currently offer ready-made entity extraction and sentiment analysis models in English, German, French, Spanish, Bulgarian, Italian, traditional Chinese, Portuguese, Russian, Swedish and Dutch.
Our solutions are immediately deployable via API, easily customisable, and can be augmented via our Data science as a service (DSaaS) capabilities. We can add entity categories and languages. Further customisation is available on demand.
By improving your services offering with the help of our AI services, you can easily identify and attract new clients for more ambitious and complex projects. Do not hesitate to make a real difference in your field of expertise.
We ensure that the content you process is comprehensive, relevant and ordered, thus helping you get basic quantitative analyses out of the way to focus on higher-end, insight-packed qualitative research.
Variety of applicable AI models that help you improve your products and expand your offering.
Our award-winning abstractive summarisation solution is tuned to understand and describe the core idea of a news article, or the essence of business, economic, crime or political events. By implementing an abstractive summarisation model to your processes, you can solve copywriting issues and benefit from production costs optimization as other our customers did.
Entity extraction, also known as named entity recognition, and is an information extraction technique, whereby key elements from text are identified and classified into predefined categories.
Sentiment analysis, also known as opinion mining or emotion AI, systematically identifies the tonality of a text, most often classifying it as positive, neutral or negative. It is widely used by businesses to analyse news content, customer reviews, survey responses, or social media conversations, among others, to determine public perceptions, to better tailor their products, services and messaging.
It shows the leading topics from different sources of news articles for the day. Document collection is clustered into groups with a similar topic which is described by a set of keywords. Each document is annotated with a quantity of comments, shares and reactions (engagement), so that engagement can be easily assigned to a particular topic. Topic modelling, a key part of the whole analysis, is based on a statistical approach and is language-agnostic.
Share of voice
Our Share of voice shows every single mention of your brand. We focus only on the relevant coverage of predefined keywords, so you can monitor not only your brand but conduct competitor analysis or gain any kind of information you need. This solution can be additionally enriched with automated sentiment analysis for greater precision.
Our automated reports save you time and effort. This solution provides you with an automated quantitative and qualitative analysis of the data you want to gain knowledge from.
With our customisation and DSaaS offering, we don’t just provide a solution – we provide an entire technological ecosystem. Lead by the belief that everyone should have access to meaningful data, our customisation and DSaaS services offer analytics and consulting, giving you our resources, technologies and data scientists. Our goal is to provide companies and organisations without internal resources for AI and analytics with a custom-built and complete data science package.
Entity extraction – a case study
A global communications agency
Calculating share of voice within articles for a beer manufacturer and seven of its brands
We extracted the names, name variations, and all expressions referring to them in the text
(such as pronouns, descriptive terms, by-names, etc.), applying a share scale and rule-based
metrics evaluation to determine what percentage of an article is dedicated to the company and its brands. As compared with human analysis, this solution is far more objective and accurate, and, of course, quicker. It is easily adaptable to any brand, group of brands, and competition analysis.
Sentiment analysis – a case study
A Data Pro
Streamlining media analysis processes internally
We worked with A Data Pro’s media analysts to train our general sentiment model, which they implemented in-house to reduce the hours they spent manually processing articles, as well as
to root out disparities between analysts. Together with our entity extraction model, this solution dramatically increased the team’s efficiency. It is currently being integrated into A Data Pro’s white label media listening platform.
AI and ML will help you stay on track.
With information ever more accessible and easier to disseminate, and every internet user able to voice their opinion to an ever wider public, keeping a close and constant tab on what people care about, how people feel, what people share is ever more crucial. We work with media intelligence providers and executives to help identify the issues and attitudes that can impact their clients and business, and offer ever broader and deeper analyses.
Our solutions can be put towards:
- Processing large volumes of coverage
- Identifying patterns and trends
- To facilitate analyses
- Assessing share of voice
- Streamlining media listening platforms
- Market research
- Coping attitudes at early product development stages
- Measuring the effect of communications campaigns
- Tailoring messaging and campaigns to specific publics
- Keeping abreast of attitudes to produce engaging content
- Upgrading media listening platform-tools