How do I use the Syllabus Assistant?

All Leganto customers can now use the Syllabus Assistant. Set the parameter  leganto_import_syllabus (Configuration >  Leganto> Features > Leganto Features) to true to enable the feature. Once enabled, library staff (not instructors) will see the Syllabus option in Leganto under Create list > Import list.

For more information watch Leganto AI Syllabus Assistant (Beta) - April 2025

How does the new Syllabus Assistant work?

The Syllabus Assistant uses large language models to analyze a file that contains references, potentially alongside other information. The model will identify references to course materials within the file. Once course materials are identified, the model sends the course materials to Leganto to create a reading list. When available, the model also sends the list or course name, and the sections in which the materials appear, resulting in a structured list in Leganto. If available, the citations may contain a note for students, such as “Read chapter 4”. If the file does not include a title for the list or course, the model will create a title for the list based on the topic of the materials.

Are you using my data to train your AI?

No. Your data is not used to directly or indirectly train LLMs. This feature utilizes pre-trained large language models (LLM) to process your content and create a list of resources. Your input is not stored by the large language model or used for any other purpose than to build the list. See the Clarivate Academia & Government - Use of Generative AI page for more information about our use of AI.

Is there a charge to use this feature?

They Syllabus Assistant is included with your Leganto subscription.

Which LLM are you using?

The Leganto team tested the feature with different LLM models to evaluate accuracy, price and speed. Because different models have different strengths, the feature uses a combination of LLMs. The feature is designed such that we can utilize different models in the future as they become available. For now, we are prioritizing accuracy.

What types of files are supported?

Leganto supports ingesting files in Word, PDF and .txt format. Leganto also supports pasting references, which are processed by the LLM in the same way as files.

Do my citations need to be in a specific citation format?

No. The AI recognizes citations in any citation format and can also recognize partial references (such as Henderson, Chapter 3). The more complete the citation, the more likely it is to link to a resource in the library collection.

What are the limitations of the feature?

We tried to resolve as many limitations as possible during the initial development of the feature, but we do not know all the limitations of the syllabus to reading list feature. Below are some of the limitations we encountered during testing:

• Some syllabi cannot be processed because they trigger content filters built into the LLMs. For more information about content filtering see the Content Filtering article from Azure OpenAI
• Some syllabi cannot be processed, and we don’t yet know why. Please share examples that don’t work at all so we can analyze them.
• Some data in tables cannot be processed because the structure of the table interferes with the structure of the text.

What languages are supported?

Most of our testing was using English language syllabi, although we did some testing in other languages. The syllabus assistant is currently removing diacritics, but we will fix this in a future release. Otherwise, any language limitations are due to limitations in the LLMs.

Why is the import slow?

For now, we have chosen to prioritize accuracy over speed. The longer your file and the more citations it contains, the longer the import process will take.

The reading list in Leganto has a mistake.

The import process is not 100% accurate, and we suggest you check the results. We are evaluating the results and expect to modify the feature in the future. We also expect to see improvements to the feature as LLMs continue to advance.

Some items didn’t link to the library where they should have.

The feature is designed to link whenever there is a match on ISBN and if no match is found, it will look for a match on title/author/year. Additionally, citations containing a DOI should match and if not, it will look for a match by ISSN, volume, issue, start page, and publication year.

It is very difficult to identify when Leganto has linked a citation to the wrong item in the library collection. Therefore, we have determined that it is better not to link to the library when we can’t be certain it is the right item. We suggest using the “Manage link to library option” to manually link items where we cannot match them automatically.