Skip to Main Content Header

Databases

ProQuest One Business

ProQuest One Business combines multiple ProQuest business information databases to a total of more than 130 million documents.

  VISIT ProQuest One Business

ProQuest One Business combines multiple ProQuest business information databases to a total of more than 130 million documents.

To facilitate the research of e.g. company or market information ProQuest developed new corporation and business subject thesauri for ProQuest One Business as well as new indexing for e.g. different types of reports and report content. Based on this new indexing ProQuest One

ProQuest One Business offers special search and navigation features that are not available elsewhere on the ProQuest Platform. 
Though it is in many ways unique ProQuest One Business resides on the ProQuest Platform and many basic functions – like navigation, basic search, personalization etc. – are the same across the whole platform.  

Q: What is ProQuest Research Assistant?
A: These new features harness AI’s powerful capabilities and applies them in a responsible, reliable manner as a research companion for students, allowing them to:

  • Easily craft more effective and targeted searches
  • More effectively review, analyze, and interrogate documents
  • Quickly evaluate the usefulness of each document for your research
  • Get guidance on next steps like choosing a research topic and understanding important concepts in the document

Q: How can the ProQuest Research Assistant help me with my search?
A: The ProQuest Research Assistant can help users craft better, more targeted searches. On the results page, it will suggest additional related terms that you can add to your initial query in order to reduce the number of results and get only to the most relevant ones. Adding additional terms will narrow the search criteria and number of results even further.
It also helps users identify terms that are related to their research topic, which helps them better understand the wider subject.

Q: What are the key takeaways and how are they generated?
A: Key takeaways are extracted from the document and are intended to help a user understand what this document is about in a succinct, easy-to-understand way to quickly determine its relevance for their specific research. The information provided is based on the actual document, and we are not using any other documents on the database or from any public LLMs to generate this. If you run a search on ProQuest, the Key Takeaways will be influenced by the search that you run to make them more relevant to you. If you run a search in ProQuest the Key takeaway section will be made up of 3 elements:
1.    A short paragraph on what the document is about.
2.    A paragraph that mentions additional topics addressed within the document.
3.    A paragraph explaining the relationship between the document and your search.

If the user arrived at the document directly without running a search – e.g. because they followed a link from an alert, discovery service, etc. – then the third paragraph would not display. 



Q: What are the important concepts and how are they generated?
A: Important concepts and their explanations are also extracted from the document. This is a list of concepts that are relevant in the context of the document, sorted in the order of importance in relation to the document. 

  • Concepts are not just definitions of the terms, they include an explanation of why the term is relevant to the document. 
  • The definition of the terms is not derived from external sources but is generated using the text of the document with the goal of explaining the concept but always in the context of the document text itself.
  • Each concept is hyperlinked so that users can run a search on any individual term by clicking on it.
  • Users can also click on ‘Find sources about these concepts’ at the bottom to run a search for all terms, combined with the OR Boolean operator.




Q: What are the research topics and how are they generated?
A: Research Topics are also generated based on the document. This feature suggests potential topics that the user might be interested in to expand their knowledge about the subject or find inspiration on which topic they want to write about. 

  • Research topics are extracted exclusively from the text of the document.
  • They’re made up of 2 components: a main topic and a more defined sub-topic.
  • Next to each research topic you will find a magnifying glass icon that will allow you to run a search to retrieve further documents related to that specific research topic.



Q: On which documents will the key takeaways, important concepts, and research topics appear?
A: These features will appear in the Research Assistant sidebar on the following content types:

  • Scholarly journals
  • Trade journals
  • Magazines
  • Conference proceedings
  • News (current and historical, including wire feeds)
  • Encyclopedia/reference works
  • Government & official publications
  • Reports
  • Working papers
  • Standards and practice guidelines

Please, note that documents under 150 words will not display these features. The features will also not appear on documents from certain publishers that have decided not to participate.

Q: How does the Research Assistant select the search terms that appear underneath the search box on the results page?
A: The ProQuest content has been tagged with specific terms belonging to Clarivate's AI-based Topic Model, and these are the terms that the Research Assistant displays on the results page. The advantage of using a Topic Model is that Topic models can dynamically adapt to new data, identifying emerging themes and trends without needing manual updates. In contrast, static taxonomies require constant maintenance to stay relevant.

The Research Assistant selects the terms or phrases from the retrieved documents that stand out because they appear more frequently in your specific results set compared to the rest of the ProQuest platform content. It does so using "ranks": every term gets 2 ranks: a search rank, which is the frequency of the term within your search results set, and a global rank, which is the frequency of the term across the entire ProQuest content. We therefore assign the term with the highest count rank 1, the next highest rank 2, etc. and we do that both at the search level and at the global level. The final score of a term/phrase that determines which terms display in the Research Assistant is just the difference between the global rank and the search rank. 

For example, a search on malaria would retrieve documents that have been tagged with the terms temperature and anopheles. If we look at how many times those terms appear across the entire ProQuest content vs. how many times they appear in this particular search results set, we see that temperature is very common across the entire ProQuest content, whereas the term anopheles seems to appear more often when you are searching on the term malaria.
           

Term Global rank Search rank Final Score
(global rank - search rank)
temperature 20 93 -73
anopheles 68 4 64


The higher score for anopheles tells us that anopheles is a much better suggestion than temperature when the user is interested in malaria. Indeed, anopheles turns out to be the genus of mosquito that transmits malaria to humans. Temperature appears on some of these search results, but it is not something that really is that unique to documents about malaria.

Q: What languages does the Research Assistant support?
A: The Research Assistant will always display text in English – with some exceptions. In non-English documents, the first paragraph is sometimes displayed in the language of the text, but the rest of the Research Assistant features display only in English.
Q: Does Research Assistant produce inaccurate statements, bias, or “hallucinations”?
A: AI-generated content has the risk of containing inaccuracies, so it’s strongly recommended that users always verify all facts. However, the way we have implemented the ProQuest Research Assistant, using Retrieval-Augmented Generation techniques, minimizes the chances of hallucinations as the information displayed is derived from the full text of the document the user is looking at. The Research Assistant will also use the same perspective as the original document used, minimizing the risk of bias.
Q: The Research Assistant didn’t extract key takeaways, important concepts nor research topics for my document, although it should. Why is that?
A: The LLM model we are using has some built-in guardrails in place to ensure that no offensive or inappropriate content is generated. Some documents that contain sensitive terms might be blocked by the LLM to ensure no harmful responses are generated.
Q: What safeguards do you have in place to protect user’s privacy?

A: At Clarivate, we are committed to upholding the highest standards in data privacy and security, especially when employing advanced AI technologies like generative AI or genAI. For an overview of our commitment with respect to your data and other data used by our tools, including interactions with our systems, please visit https://clarivate.com/ai/academia/policy/

Q: Is the ProQuest Research Assistant accessible?
A: We have designed and built Research Assistant to be compliant with WCAG 2.2 guidelines (levels A and AA). This means, for example, that users should be able to use Research Assistant with a screen reader, with keyboard only navigation, and with devices of various sizes and zoom levels. If you find any issues with the accessibility of Research Assistant, please contact us.

Q: I don’t want my users to see the ProQuest Research Assistant. How can I turn it off?
A: Users can hide the Research Assistant anytime by using the arrow icon on the top right corner of the feature. If you would like to disable the feature for all users on your account, you can contact our support team . This means that end users cannot switch the tool on and off during their session.
However, we strongly recommend that you keep the ProQuest Research Assistant activated on your account because:
 
1.    End users want to use genAI tools and the ProQuest Research Assistant is a safe environment to expose your users to responsible genAI tools, with all security and privacy considerations taken into account – as opposed to tools like ChatGPT.
2.    The Research Assistant’s goal is not to do work for the end user but support them in their research task instead.
3.    On the results page: 

  • It helps educate users on how to easily craft more effective and targeted searches.
  • It helps users get to a smaller set of results that is more relevant and more specific.

4.    On the document side: 

  • It empowers users to review, analyse, and interrogate documents more efficiently.
  • It helps users evaluate the usefulness of each document for their research, saving time and boosting productivity.
  • It allows the user to move their research on to exciting new research areas.
  • It drives student engagement and higher usage.

Using publications in ProQuest One Business

Proquest Basic Search

footer