The webAI SDG 3 Agent was trained by ISTARI.AI to derive companies’ commitment to Sustainable Development Goal 3 (Good Health and Well-Being) from their websites and to display it as an individual SDG 3 Intensity Score. This numerical indicator reflects how centrally the topic of Sustainable Development Goal 3 is communicated by the company on its own website and presented as essential for its own business model.

SDG 3 is to be understood as a broad concept. This includes services and products that are used directly to improve health and well-being (e.g., pharmaceutical companies, healthcare providers, or manufacturers of medical equipment), but also, for example, businesses involved in health education, mental health services, and wellness initiatives. Thus, players from sectors such as healthcare, biotechnology, fitness, and mental health care are considered.

Moreover, WebAI also identifies companies that specialize in developing health technologies, such as telemedicine or wearable health devices, organizations that focus on health-related research, or those involved in providing access to clean water and sanitation, which are crucial for maintaining good health.

The webAI SDG 2 Agent was trained by ISTARI.AI to derive companies’ commitment to Sustainable Development Goal 2 (Zero Hunger) from their websites and to display it as an individual SDG 2 Intensity Score. This numerical indicator reflects how centrally the topic of Sustainable Development Goal 2 is communicated by the company on its own website and presented as essential for its own business model.

SDG 2 is to be understood as a broad concept. This includes services and products that are used directly to combat hunger and ensure food security (e.g., agricultural technology companies or food distribution organizations), but also, for example, suppliers of sustainable farming equipment, and companies involved in food preservation and waste reduction. Thus, among others, players from the agriculture, food production, distribution, and technology sectors are considered.

However, WebAI also finds, for example, companies that specialize in developing innovative food solutions, such as alternative protein sources, organizations that focus on agricultural education and research, or businesses that create efficient supply chain solutions to reduce food loss and improve access to nutritious food.

Our webAI reads the website of the company under investigation and searches for text sections (paragraphs) that deal with the topic of SDG 2. To do this, webAI first searches for keywords that are potentially related to SDG 2, analyzes the paragraphs identified in this way, and determines whether or not they are actually relevant texts. If webAI has assigned a corresponding paragraph to the topic of SDG 2 with a high probability, webAI remembers this paragraph and continues searching. In this way, webAI searches the entire company website or the particularly relevant “top-level” sub-pages if it is a very extensive website with hundreds of sub-pages.

In this way, webAI finds a certain number of paragraphs per company website that deal with the topic of SDG 2. WebAI then puts this number in relation to the total amount of text content read on the website. In this way, webAI determines a SDG 2 intensity for each company.

The SDG 2 Intensity Score calculated in this way would be 0.0 for a company with no SDG 2-related texts. For a consulting company, for example, the value could be 0.18, and for a startup that is particularly focused on SDG 2-related topics, it could be 3.05. The SDG 2 Intensity Score has no upper limit.

In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the intensities from “low” to “very high”. In the example above, the consulting company would be classified as “low” intensity and the startup as “very high” intensity.

Unlike simpler, binary classifications (“SDG 2 YES/NO”), webAI therefore outputs a continuous score at company level. This makes it possible to distinguish between companies for which SDG 2-related topics are only a marginal issue and those for which they play a central role. Users of webAI data can thus easily determine for themselves how strongly a company should focus on SDG 2 in order for it to be relevant to them.

Like all our webAI agents, the webAI SDG2 Agent was developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then delivers expert-level results. The SDG2 Intensity Scores are currently also being validated in a not yet published study.

The webAI Additive Manufacturing Agent (3D Printing) was trained by ISTARI.AI to derive companies’ know-how in the field of additive manufacturing from their websites and to map it as an individual Additive Manufacturing Score. By knowhow in this context, we mean products and services in additive manufacturing or personnel with skills in additive manufacturing. This numerical indicator reflects how central the topic of additive manufacturing is communicated by the company on its own website and presented as essential for its own business model.

In addition, webAI also derives an Additive Manufacturing Information Intensity Score for each company. This reflects how intensively the company communicates about the topic of additive manufacturing without having its own products and services with integrated additive manufacturing or specially trained personnel.

Companies that are active in the field of additive manufacturing, have business areas geared to it, or offer products and services with a direct link, usually communicate this fact. The more central this topic is for the company, the more significant it is for the company’s external communication. For example, a startup for “rapid prototyping” communicates almost exclusively on the topic of additive manufacturing, while a consulting company that offers “additive manufacturing consulting” among many other topics communicates only to a limited extent on additive manufacturing. WebAI distinguishes “information” communication in addition to this “knowhow” communication (it offers its own additive manufacturing products and services). An example of this would be a regional newspaper’s website reporting that a regional incubator for additive manufacturing startups has recently opened.

Our webAI reads the website of the company under investigation and searches it for text sections (paragraphs) that deal with the topic of additive manufacturing. For this purpose, webAI first searches for keywords related to additive manufacturing, then analyzes the identified paragraphs and determines whether the company reports on its own additive manufacturing know-how or only communicates information about additive manufacturing. If webAI has assigned a corresponding paragraph to the category “Knowhow” or “Information”, webAI remembers this paragraph and continues searching. In this way, webAI searches the entire company website, or the particularly relevant “top-level” sub-webpages in the case of very extensive websites with hundreds of sub-webpages (for more information, see the publication: ). WebAI thus finds a certain number of paragraphs per company website that deal with the topic of additive manufacturing. WebAI classifies these paragraphs into “knowhow” and “information” and then relates the number to the total amount of text content read on the website. Thus, webAI determine an Additive Manufacturing Knowhow Intensity and an Additive Manufacturing Information Intensity for the company.

The intensities determined in this way would be 0.0 for a company with no additive manufacturing-related texts. For the example of a consulting company described above, on the other hand, the value for Additive Manufacturing Knowhow could be 0.25 and the value for Additive Manufacturing Information 0.21. The startup that is particularly focused on additive manufacturing could have an Additive Manufacturing Knowhow Intensity of 3.8 and an Additive Manufacturing Information Intensity of 0.9. The regional newspaper, on the other hand, would have an Additive Manufacturing Knowhow Intensity of 0.0 and an Additive Manufacturing Information Intensity of 0.05.

In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the intensities from “low” to “very high”. In the example above, the consulting company would be classified as “low” intensity and the startup as “very high” intensity.

So, unlike simpler, binary classifications (“additive manufacturing YES/NO”), webAI outputs two continuous scores at the company level. This allows to distinguish between companies where additive manufacturing is only a marginal topic and those where it plays a central role. In addition, companies with products and services based on or related to additive manufacturing can be distinguished from those that merely provide information on the topic. Users of webAI data can thus easily determine for themselves how and to what extent a company should focus on additive manufacturing so that it is relevant to them.

Like all our webAI agents, the webAI Additive Manufacturing Agent was developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then generates expert-level results. For this agent, ISTARI.AI collaborated with the Technical University of Munich.

The validation of the Additive Manufacturing Intensity Scores was also carried out in the study , which was conducted by researchers from the University of Mannheim, University of Giessen, University of Salzburg and the Technical University of Munich.

The webAI Standards Agent has been trained by ISTARI.AI to search companies’ websites for standards and certifications. These are standards from the International Organization for Standardization (ISO), the International Electrotechnical Commission (IEC) and national standardization organizations such as the German Institute for Standardization (DIN). Well-known and widely used standards include paper formats (ISO 216, DIN 476) such as DIN A4, the quality management standards of the ISO 9000 family (e.g. ISO 9001) or the information security management standard ISO/IEC 27001.

Our webAI reads the website of the company under investigation and searches it for mentions of ISO, IEC and standards issued by other organizations. For this purpose, webAI examines the entire company website, or the particularly relevant “top-level” sub-webpages if the website is very large and contains hundreds of sub-webpages (for more information, see the publication: ). WebAI thus finds a certain number of standard mentions per company website and returns them as a list (e.g. [DIN EN ISO 9001, DIN EN ISO 50001, ISO 9001:2…]).

From the lists of found standards generated by the webAI Standards Agent, specific standards (e.g. “ISO 9001:2000”) can be determined as well as the number of mentions of a specific standard (e.g. “‘DIN EN 16247-1’: 68, ‘DIN EN ISO 50001’: 34”). As an example for the DACH region (Germany, Austria, Switzerland), it can be said that about 10% of the companies surveyed mention at least one standard on their website. On average, these companies mention one standard 8.7 times (median 2.0) and different standards 2.6 times (median 1.0). The most frequently listed standards in the DACH region are, by a wide margin, ISO 9001, followed by ISO 14001 and ISO 27001.

It is important to note here that our approach does not differentiate whether companies are actually certified or name a standard for other reasons (e.g., because they work with certified companies). Specifically for the ISO/IEC 27001 standard, a detailed manual review (for more information, see the publication: ) of webAI identified companies yielded the following result regarding the reason for citation:

Like all of our webAI agents, the webAI Delivery Delay Agent has been developed and validated in conjunction with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then generates expert-level results. For this agent, ISTARI.AI collaborated with researchers from the Technical University of Berlin. Among other things, the generated data was used for the study , which was published in the journal IEEE Transactions on Engineering Management.

The webAI Company Networks Agent captures hyperlinks between companies’ websites to map their interconnectedness. Companies link to the websites of other economic actors for various reasons. For example, companies often name reference customers on their websites and then usually also set up a hyperlink to their website. With webAI we quantify these linkages by identifying hyperlink connections between companies.

Our webAI reads the website of the examined company and searches it for hyperlinks that target other websites (domains). Thus, webAI searches the entire corporate website, or the particularly relevant “top-level” sub-webpages, if very extensive websites with hundreds of sub-webpages are involved (for more information, see the publication: ).

WebAI thus finds a certain number of hyperlinks per company website that point to external websites. Then, webAI merges this information with content from all other corporate websites that were also searched. Thus, webAI identifies not only “outbound” hyperlinks, i.e., those pointing from Company A to Companies B, C, and D, but also “inbound” hyperlinks, i.e., hyperlinks from Companies B, E, and F pointing to Company A. From this information, webAI calculates the following variables that provide information about the interconnectedness of the company:

1. Incoming links: The destination domains of all outbound hyperlinks from a company’s website.

The data collected with webAI Company Networks maps the hyperlink connections between the examined company websites. This relational data can be examined for more complex analysis using specialized software for network analysis. For example, centrality measures or cluster detection algorithms can be applied. However, the data provided also allows valuable insights into the networking of individual companies without further analysis.

For example, a glance at the “Links Count” column reveals how many other companies the company under consideration is connected to via a hyperlink. This variable combines both all outgoing links from the company under consideration and all incoming links from other companies. As an example for the DACH region (Germany, Austria, Switzerland), it can be said that companies have an average “Links Count” of 13.1 (median 6.0), with 9.4% of companies having no hyperlinks to other companies at all. The maximum achieved value for “Links Count” is 559,203, which is google.com. Depending on the nature of the question, it is recommended to remove such outliers from the data.

Like all our webAI agents, the webAI Comapny Networks was developed and validated together with independent subject matter experts. This way, we ensure that the results of the webAI agents are validated in the context of external studies at a high scientific level. For the development of this agent ISTARI.AI collaborated with ZEW – Leibniz Centre for European Economic Research in Mannheim, the University of Salzburg and the Technical University of Berlin. In addition, the results of the agent have been used in scientific studies, including by researchers at the University of Groningen in the Netherlands and in a joint study with researchers at ETH Zurich. The latter study was awarded the Best Paper Award at the renowned “R&D Management Conference 2022”.

The webAI Blockchain Agent was trained by ISTARI.AI to derive companies’ Blockchain know-how from their websites and to map it as an individual Blockchain Intensity Score. Blockchain know-how in this context is products and services with integrated Blockchain or personnel with Blockchain skills. This numerical indicator reflects how central the topic of Blockchain is communicated by the company on its own website and presented as essential for its own business model.

Companies that are active in the area of Blockchain, have business fields geared to it or offer products and services with a direct connection usually communicate this fact. The more central this topic is for the company, the more significant it is for the company’s external communication. For example, a startup for integrating blockchain into the supply chain communicates almost exclusively on the topic of blockchain, while a consulting company that offers “blockchain consulting” among many other topics communicates only to a limited extent about this technology. WebAI distinguishes “information” communication in addition to this “know-how” communication (it offers its own products and services with integrated Blockchain). An example of this would be the website of a regional newspaper reporting that a regional incubator for Blockchain startups has recently opened.

Our webAI reads the website of the company under investigation and searches it for text sections (paragraphs) that deal with the topic of blockchain. For this purpose, webAI first searches for keywords related to blockchain, then analyzes the paragraphs identified in this way and determines whether the company reports on its own blockchain know-how or merely communicates information on the topic of blockchain. If webAI has assigned a corresponding paragraph to the category “Know-how”, webAI remembers this paragraph and continues searching. In this way, webAI searches the entire company website, or the particularly relevant “top-level” sub-webpages, if very extensive websites with hundreds of sub-webpages are involved (for more information, see the publication: ).

WebAI thus finds a certain number of paragraphs per corporate website that deal with the topic of blockchain. WebAI classifies identifies “know-how” paragraphs and then relates their number to the total amount of text content read on the website. Thus, webAI determine a Blockchain Intensity for the company.

The intensity determined in this way would be 0.0 for a company with no Blockchain-related texts. For the example of a consulting company described above, on the other hand, the value for Blockchain Intensity could be 0.25. The startup that is particularly focused on Blockchain could have a Blockchain Intensity of 3.8. The regional newspaper, on the other hand, would have a Blockchain Intensity of 0.0.

In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the intensities from “low” to “very high”. In the example above, the consulting company would be classified as “low” intensity and the startup as “very high” intensity.

So, unlike simpler binary classifications (“Blockchain YES/NO”), webAI outputs two continuous scores at the company level. This allows us to distinguish between companies where Blockchain is only a marginal topic and those where it plays a central role. Users of webAI data can thus easily determine for themselves how and to what extent a company should focus on Blockchain so that it is relevant to them.

As an example for the DACH region (Germany, Austria, Switzerland), it can be said that 0.5% of the companies surveyed have a Blockchain Intensity greater than 0.0 and thus communicate Blockchain know-how in some form. The average Intensity Score for those companies is 0.45, with half of the companies having a score greater than 0.18. The maximum score achieved is 4.2, with 14.0% achieving a score of 1.0 or higher.

Like all our webAI agents, the webAI Blockchain Agent has been developed and validated together with independent domain experts. This ensures that we train the webAI agent with real expert knowledge and that it then generates expert-level results. For this agent, ISTARI.AI collaborated with the Technical University of Munich.

The validation of the Blockchain Intensity Scores is also currently being carried out in a study with the working title “Blockchain Technology Diffusion and Use Cases in the European Industry”, which is being conducted by researchers from the University of Mannheim, University of Giessen and the Technical University of Munich.

The webAI Social Innovation agent was trained by ISTARI.AI to derive from companies’ websites their engagement in social innovation and to map it as an individual Social InnoProb Score. This indicator, ranging from 0.0 to 1.0, reflects the likelihood that a company is a social innovator.

Social innovations are innovations that improve human interaction in many areas, for example, how people work, organize their leisure time, shop, live or how they are mobile. However, unlike product innovations, there is still no internationally standardized definition of social innovations. We at ISTARI.AI follow the current working definition as it is currently being elaborated by OECD and Eurostat: Social innovation also described as the emergence, implementation and diffusion of new social practices in the societal domain that are directly related to the search for viable and sustainable solutions to societal problems and challenges. If a company has introduced such a social innovation in the last three years through changes in the company or brought it to market as a new or improved product or service, then it is considered a social innovator.

Websites are used by companies as platforms to provide information about their products and services, performance, strategies and relationships. All these aspects can be associated with the social innovations developed by the company. The Social InnoProb Score is not intended to identify individual innovations (e.g. new products), but companies with an innovation-related activity profile. This probability can be interpreted as a continuous innovation indicator at the company level.

We use a traditional company-level indicator based on a questionnaire-based innovation survey (Community Innovation Survey Germany) to train webAI with the websites of socially innovative and non-innovative companies. In doing so, we use survey waves from several years to identify companies that answered positively to questions related to social innovation. Examples of such questions are the cooperation of the company with associations, the introduction of new methods of work organization or the promotion of togetherness and family atmosphere in the company. During this training webAI learns independently how the websites of socially innovative and non-innovative companies differ. After the training, we can then apply webAI to any other company website. Based on the textual content, webAI then calculates how likely it is that the company under investigation is innovative.

The Social InnoProb score calculated in this way would be close to 0.0 for a company that is very unlikely to be a socially innovative company. A very likely socially innovative company, on the other hand, would have a Social InnoProb score close to 1.0.

In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the probabilities of being an innovative company from “very low” to “very high”.

Thus, unlike simpler, binary classifications (“innovative YES/NO”), webAI outputs a continuous score at the company level. Users of webAI data can thus easily determine for themselves the probability with which a company should be socially innovative in order to be relevant to them. For the binary classification into socially innovative and non-innovative companies, we recommend a classification threshold of 0.5.

As an example for the DACH region (Germany, Austria, Switzerland), we can say that 24.35% of the analyzed companies are rated as socially innovative by webAI Social InnoProb (Social InnoProb Score > 0.5). The average Social InnoProb score of all companies is 0.44. Half of the companies have a score of more than 0.41. The highest score achieved is 0.85 and the lowest is 0.14. The figure below shows the distribution of the Social InnoProb score for the DACH region.

Like all our webAI agents, the webAI Social InnoProb agent was developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it subsequently generates expert-level results. For this agent, ISTARI.AI collaborated with . The ZEW is one of the leading German economic research institutes with a high European reputation.

The webAI InnoProb agent was trained by ISTARI.AI to derive the status of companies as innovators from their websites and to map it as an individual Innovation Probability Score (InnoProb). This indicator between 0.0 and 1.0 reflects the probability of a company being an innovator.

Innovation is defined here in terms of the OECD Oslo Manual: An innovation is the introduction of a new or significantly improved product (good or service) or process, a new marketing method, or a new organizational method in business practice, workplace organization, or external relations.

Websites are used by companies as platforms to provide information about their products and services, performance, strategies, and relationships. All these aspects can be associated with the innovations developed by the company. The InnoProb Score does not intend to identify individual innovations (e.g. new products), but companies with an innovation-related activity profile. This probability (the InnoProb Score) can be interpreted as a continuous innovation indicator at company level.

We use a traditional company-level indicator based on a questionnaire-based innovation survey (Community Innovation Survey for Germany) to train webAI with the websites of innovative and non-innovative companies. In this training webAI learns independently how the websites of innovative and non-innovative companies differ. After the training, we can then apply webAI to any other company website. Based on the textual content, webAI then calculates how likely it is that the company under study is innovative.

The InnoProb score calculated in this way would be close to 0.0 for a company that is highly unlikely to be an innovative company. A very likely innovative company, on the other hand, would have an InnoProb Score close to 1.0.

In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the probabilities of being an innovative company from “very low” to “very high”.

So, unlike simpler, binary classifications (“innovative YES/NO”), webAI outputs a continuous score at the company level. Users of webAI data can thus easily determine for themselves the probability with which a company should be innovative in order to be relevant to them. For the binary classification into innovative and non-innovative companies, we recommend a classification threshold of 0.4.

As an example for Germany, we can say that 15.12% of the examined companies are assessed as innovative by webAI InnoProb (InnoProb Score > 0.4). The average InnoProb Score of all companies is 0.25. Half of the companies have a score greater than 0.20. The maximum score achieved is 0.94 and the lowest is 0.03. The figure below shows the distribution of InnoProb scores as a histogram.

Like all our webAI agents, the webAI InnoProb agent has been developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then generates expert-level results. For this agent, ISTARI.AI collaborated with the University of Giessen and the ZEW – Leibniz Centre for European Economic Research in Mannheim. ZEW is one of the leading German economic research institutes with a high European reputation.

The InnoProb scores were also validated by comparing them with patent data from the European Patent Office, with projections based on survey data from the Mannheim and Berlin Innovation Panels, and with regional innovation indicators from the German Federal Statistical Office. The corresponding study was published in the international journal PLOS ONE.

The webAI Mobility Agent was trained by ISTARI.AI to derive companies’ commitment to mobility from their websites and to display it as an individual Mobility Intensity Score. This numerical indicator reflects how centrally the topic of mobility is communicated by the company on its own website and presented as essential for its own business model.

Mobility is to be understood as a broad concept. This includes services and products that are used directly as a means of transport (e.g. manufacturers of e-bikes or automobiles), but also, for example, suppliers of components and software for the production of means of transport or mobility service providers such as public transport. Thus, among others, players from the automotive industry, railroad, aerospace, shipping, logistics, telematics (e.g. intelligent transport systems), e-mobility or fuel cell technology are considered. However, WebAI also finds, for example, companies that specialize in the construction of underground garages, electricians who install wallboxes, or chemical manufacturers who produce lubricants for the automotive industry.

Companies that are active in these mobility-related areas, have business fields geared to them, or offer products and services with a direct connection, usually communicate this on their websites. The more central the topic is for the company, the more significant it is for the company’s external communication. For example, a startup for bicycle helmets communicates almost exclusively on the topic of mobility, while a chemical company, which among many other products also offers specific soaps for automotive workshops, does so only to a limited extent.

Our webAI reads the website of the company under investigation and searches for text sections (paragraphs) that deal with the topic of mobility. To do this, webAI first searches for keywords that are potentially related to mobility, analyzes the paragraphs identified in this way, and determines whether or not they are actually mobility-related texts. If webAI has assigned a corresponding paragraph to the topic of mobility with a high probability, webAI remembers this paragraph and continues searching. In this way, webAI searches the entire company website or the particularly relevant “top-level” sub-pages if it is a very extensive website with hundreds of sub-pages (for more information, see the publication: ).

In this way, webAI finds a certain number of paragraphs per company website that deal with the topic of mobility. WebAI then puts this number in relation to the total amount of text content read on the website. In this way, webAI determines a mobility intensity for each company.

The Mobility Intensity Score calculated in this way would be 0.0 for a company with no mobility-related texts. For the example of a chemical company described above, on the other hand, the value could be 0.15, and 3.97 for the startup that is particularly focused on mobility. The Mobility Intensity Score has no upper limit. In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the intensities from “low” to “very high”. In the example above, the chemical company would be classified as “low” intensity and the startup as “very high” intensity.

In contrast to simpler, binary classifications (“Mobility YES/NO”), webAI outputs a continuous score at the company level. This makes it possible to distinguish between companies for which mobility is only a peripheral issue and those for which it plays a central role. Users of the webAI data can thus easily determine for themselves how strongly a company should adopt mobility for it to be relevant to them.

As an example for the DACH region (Germany, Austria, Switzerland), it can be said that 20.5% of the companies studied have a Mobility Intensity Score of greater than 0.0 and thus communicate the topic of mobility in some form on their websites. The average Mobility Intensity Score of these companies is 0.61 and half of the companies have a score lower than 0.30. The maximum score achieved is 6.93, with 21.63% achieving a score of 1.0 or higher.

Like all our webAI agents, the webAI Mobility Agent was developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then delivers expert-level results. For this agent ISTARI.AI has collaborated with researchers from the . The Mobility Intensity Scores are currently also being validated in a not yet published study.

The webAI Energy Agent was trained by ISTARI.AI to identify companies with a focus on the energy sector and to map their focus on this topic as an individual Energy Intensity Score. This numerical indicator measures how centrally the topic of energy is communicated by the company on its own website and presented as essential for its own business model.

We consider the entire value chain from the exploitation of energy sources to distribution to end consumers and their consumption patterns, i.e. the entire energy industry. More specifically we look at four sub-segment:

• Generation: Power/energy generation is the colloquial term for energy transformation where electrical energy and heat is made available. All companies that are involved in the exploitation of energy sources up to their transformation into electrical energy or heat belong to this category. Companies that provide products or services for this purpose also belong to this category. Examples are: Oil production (oil drilling), coal mining, operators of solar parks, manufacturers of solar panels , manufacturers of steam boilers for power generation in coal-fired power plants, manufacturers of control software for nuclear power plants, waste incineration for district heating, hydrogen production.

• Grid: Network for the transmission and distribution of electrical energy, heat and other energy sources (gas, hydrogen etc.). This includes the manufacturers of required components (hardware and software), service providers and operators. Examples are: Contractors and commissioning authorities for the development and operation of the electricity grid and pipelines (gas, district heating, etc.), manufacturer of control software for power grids (smart grid), manufacturer and operator of energy storage systems, petrol stations, electricians who install wall boxes for electric cars and manufacturers of wall boxes, manufacturers and operators of charging stations for electric cars, manufacturers of smart electricity meters, shipping company with oil tankers.

• Efficiency: Efficient use and production of energy and optimized processes to reduce energy consumption, energy loss and emissions while maintaining or increasing effectiveness (“same or more for less”). All companies that are active in this area or help other companies/consumers to be more efficient belong to this category. Examples are: Energy consultant for companies and private individuals, construction company that carries out energy refurbishments of buildings, manufacturer of thermal insulation material, companies converting their vehicle fleet to models with lower fuel consumption, company that is renovating its own company headquarters to make it more energy efficient, power producer introducing measures to reduce energy loss in power generation with coal, company switches completely to renewable energies in production.

• Market: Includes trading of energy sources and electricity, as well as the legal framework and (state) support measures. Examples are: Energy exchanges where energy carriers (e.g. gas) and electrical energy are traded, electricity suppliers who supply the end customer but do not operate any infrastructure themselves, legislative and supervisory bodies in the field of electricity and gas markets, development banks (e.g. German KfW) with programs to promote energy-efficient construction and refurbishment, portal for comparison of energy providers for private customers.

Companies that are active in these energy-related areas, have business fields geared to them, or offer products and services with a direct connection, usually communicate this on their websites. The more central the topic is for the company, the more significant it is for the company’s external communication. For example, a startup for smart grid software communicates almost exclusively on the topic of energy, while a chemical company, which among many other products also offers specific thermal insulation components, does so only to a limited extent.

Our webAI reads the website of the company under investigation and searches for text sections (paragraphs) that deal with the topic of energy. To do this, webAI first searches for keywords that are potentially related to energy, analyzes the paragraphs identified in this way, and determines whether or not they are actually energy-related texts. If webAI has assigned a corresponding paragraph to the topic of energy with a high probability, webAI remembers this paragraph and continues searching. In this way, webAI searches the entire company website or the particularly relevant “top-level” sub-pages if it is a very extensive website with hundreds of sub-pages (for more information, see the publication: ).

In this way, webAI finds a certain number of paragraphs per company website that deal with the topic of energy. WebAI then puts this number in relation to the total amount of text content read on the website. In this way, webAI determines a Energy Intensity for each company.

The Energy Intensity Score calculated in this way would be 0.0 for a company with no energy-related texts. For the example of a chemical company described above the calculated intensity value could be 0.15, and 3.97 for the startup that is particularly focused on energy. The Energy Intensity Score has no upper limit. In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the intensities from “low” to “very high”. In the example above, the chemical company would be classified as “low” intensity and the startup as “very high” intensity.

In contrast to simpler, binary classifications (“Energy YES/NO”), webAI outputs a continuous score at the company level. This makes it possible to distinguish between companies for which energy is only a peripheral issue and those for which it plays a central role. Users of the webAI data can thus easily determine for themselves how strongly a company should be engaged in the energy sector for it to be relevant to them.

As an example for the DACH region (Germany, Austria, Switzerland), it can be said that 20.5% of the companies studied have a Energy Intensity Score of greater than 0.0 and thus communicate the topic of energy in some form on their websites. The average Energy Intensity Score of these companies is 0.61 and half of the companies have a score lower than 0.30. The maximum score achieved is 6.93, with 21.63% achieving a score of 1.0 or higher.

Like all our webAI agents, the webAI Energy Agent was developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then delivers expert-level results. For this agent ISTARI.AI has collaborated with researchers from the . The Energy Intensity Scores are currently also being validated in a not yet published study.

The webAI Sustainability Agent was trained by ISTARI.AI to derive from the websites of companies their commitment in the area of sustainability and to depict it as an individual Sustainability Intensity Score. This numerical indicator reflects how centrally the topic of sustainability is communicated by the company on its own website and presented as essential for its own business model.

In this sense, sustainability is to be understood as “ecological sustainability”. This therefore refers to concepts such as circular economy, the energy transition, ecological agriculture, regenerative energy, efficient use of resources, reduction of emissions or recycling.

Companies that are active in these or related subject areas, have business fields geared to them, or offer products and services with a direct connection, usually communicate this on their websites. The more central this topic is for the company, the more significant it is for the company’s external communications. For example, a startup for compostable ballpoint pens communicates almost exclusively on the topic of sustainability, while a consulting company that offers “ESG consulting” among many other topics does so only to a limited extent.

Our webAI reads the website of the company under investigation and searches it for text sections (paragraphs) that deal with the topic of sustainability. To do this, webAI first searches for keywords potentially related to sustainability, analyzes the paragraphs identified in this way, and in doing so determines whether or not they are actually sustainability-related texts. If webAI has assigned a corresponding paragraph to the topic of sustainability with a high probability, webAI remembers this paragraph and continues searching. In this way, webAI searches the entire company website, or the particularly relevant “top-level” sub-webpages if it is a very extensive website with hundreds of sub-webpages (for more information, see the publication: ).

WebAI thus finds a certain number of paragraphs per company website that deal with the topic of sustainability. WebAI then relates this number to the total amount of text content read on the website. In this way, webAI determines a Sustainability Intensity for each company.

The Sustainability Intensity Score calculated in this way would be 0.0 for a company with no sustainability-related texts. For the example of a consulting company described above, on the other hand, the value could be 0.25, and for the startup that is particularly focused on sustainability, it could be 2.37. The Sustainability Intensity Score has no upper limit.

In addition, we integrate an auxiliary column into our data as an interpretation and reading aid. It categorizes the intensities from “low” to “very high”. In the example above, the consulting company would be classified as “low” intensity and the startup as “very high” intensity.

Unlike simpler, binary classifications (“sustainable YES/NO”), webAI therefore outputs a continuous score at company level. This makes it possible to distinguish between companies for which sustainability is only a marginal issue and those for which it plays a central role. Users of webAI data can thus easily determine for themselves how strongly a company should focus on sustainability in order for it to be relevant to them.

As an example for the DACH region (Germany, Austria, Switzerland), it can be said that 17.5% of the companies examined have a Sustainability Intensity Score of greater than 0.0 and thus communicate the topic of sustainability in some form on their websites. The average Sustainability Intensity Score of these companies is 0.53 and half of the companies have a score smaller than 0.28. The maximum score achieved is 6.94, with only 2.88% achieving a score of 1.0 or higher.

Like all our webAI agents, the webAI Sustainability Agent has been developed and validated together with independent subject matter experts. This ensures that we train the webAI agent with real expert knowledge and that it then generates expert-level results. For this agent, ISTARI.AI collaborated with the OECD’s Centre for Entrepreneurship, SMEs, Regions & Cities. The OECD is an international organization with 38 member states and nearly sixty years of experience in analyzing political, economic and social developments.

Our Sustainability Intensity Scores were also validated in a study on “” published by researchers from the University of Salzburg, University of Heidelberg, Harvard University and the University of Giessen in the prestigious journal Science of The Total Environment.

The webAI Digital Health Agent was trained by ISTARI.AI to identify companies with a focus on the digital health sector and to map their focus on this topic as an individual Digital Health Intensity Score. This numerical indicator measures how centrally the topic of digital health is communicated by the company on its own website and presented as essential for its own business model. Digital Health includes theuse of information and communication technology (ICT) in the field of healthcare. More specifically we look at four sub-segment:

• E-health: Companies that offer services or products that are used to support the treatment and care of patients with modern ICT methods. This includes, for example, the communication of medical data made available with an electronic health card. Electronic prescriptions for medicines and electronic patient records also fall into this category. Examples are: Pharmacies that accept electronic prescriptions, doctors and clinics that allow appointments to be made digitally, manufacturer of card readers for electronic health cards, developers offering software for electronic medical records or patient management systems, providers and users of software for online consultations, online pharmacies.

Yes, just click on the next section "Overview" below and we will guide you step by step through the various functions of the webAI Dashboard.

The webAI AI Agent was trained by ISTARI.AI to derive companies’ artificial intelligence (AI) know-how from their websites and to map it as an individual AI Intensity Score. By AI know-how, in this context, we mean products and services with integrated AI or personnel with AI skills. The resulting numerical indicator reflects how centrally the topic of artificial intelligence is communicated on the company’s website and presented as essential for its own business model.

Companies that are active in the field of artificial intelligence, have business areas geared to it, or offer products and services with a direct link, usually communicate this. The more central this topic is for the company, the more significant it is for the company’s external communications. For example, a machine translation startup communicates almost exclusively on the topic of artificial intelligence, while a consulting company that offers “AI consulting” among many other topics communicates only to a limited extent on artificial intelligence. An example of a company that only communicates about AI would be a regional newspaper’s website reporting that a regional incubator for AI startups has recently opened.

The webAI SDG 1 Agent was trained by ISTARI.AI to derive companies’ commitment to Sustainable Development Goal 1 (No Poverty) from their websites and to display it as an individual SDG 1 Intensity Score. This numerical indicator reflects how centrally the topic of Sustainable Development Goal 1 is communicated by the company on its own website and presented as essential for its own business model.

SDG 1 is to be understood as a broad concept. This includes services and products that are used directly to reduce poverty (e.g., microfinance institutions or affordable housing developers), but also, for example, suppliers of basic necessities and companies providing essential services or products for impoverished communities. Thus, among others, players from various industries like agriculture, education, healthcare, and basic utilities are considered.

However, WebAI also finds, for example, companies that specialize in the development of affordable technology solutions, organizations that provide financial literacy training, or manufacturers who produce affordable clothing and essential goods for low-income populations.

• webAI Dashboard: release of completely new, more individualised user interface with many novel options (e.g. modules).

• webAI Credit System: introduction of credits to perform requests.

• webAI Q&A: introduction of new interactive module to find company-specific answers to your questions.

• Subscription changes: Conversion to new product portfolio with webAI Lead and webAI Insights, including usage-based billing for downloads. Migration of all existing legacy subscriptions.

• Downloader: Introduction of the downloader for the individual compilation of export tables.

• Chat Helper: With the new update we introduce the alpha version of our webAI Chat Bot. The chat bot helps generating queries for the webAI platform. The user’s search query can be provided to the Chat Bot written in natural language and it will then return a matching KQL query that can be entered by the user in the webAI platform. The chat bot can be accessed directly from the dashboard via a new button.

• Contact data: With the new update, more contact information has been implemented. Besides the already known “main_contact_mail”, the new field “all_mails” has been added, which contains all email addresses listed on the website. Additionally, a new field “main_contact_number” has been added, which contains the main contact phone number listed on the website. Additionally, all other phone numbers that have been detected can be found in the “all_phones” column. The logic for extracting the phone numbers is a first version that will provide even better results in the future.

In addition to using our indicator to filter companies with online shops, we build an additional indicator to discard retailers. We indicate the probability of a company website to belong to a retailer from low probability to high probability. Each class can be Include or Exclude and you may reset the filters just as with the geography filters described above.

Additional filter to discard news providers (i.e. news outlets like newspapers, online magazines, academic journals, blogs). Functionality is identical to the drop-down menus described above. In some searches it can be advantageous to exclude such websites. For example, if you are looking for an emerging technology, you may only want to find manufacturing companies that already mention this technology on their website, but not media companies that report on this technology on their websites. We indicate the probability of a company website to belong be a news provider from low probability to high probability.

Additional filter to discard domain providers. Functionality is identical to the drop-down menus described above. When companies shut down business and their domains (i.e. website addresses) are up for sale, the domain vendors (i.e. domain provider) shows ads when accessing these domains. In order to filter such cases you may Exclude the 1 values using this drop-down menu.