Doc Chat
Last updated
Last updated
DocChat contains two sub-modules. Both are ideal for quickly getting to know your document.
Chat with your document allows you to interactively "interrogate" your document (and even other PDF/DOCX files) with the help of Artificial Intelligence (AI). Ask a question, and get both an concise answer and an overview of all the relevant parts of the document that deal with your question.
Semantic Search allows you to search in your document for semantically related words.
The DocChat module works similar to other Generative AI chatbots, such as ChatGPT and Microsoft CoPilot. Essentially you ask it a question, it will provide you with a response, and you can then ask follow-up questions and hold a conversation with the chat bot.
The largest difference is that ClauseBuddy's chatbot is "layout-aware". If its answer consists of texts that you will likely want to insert into your opened MS Word document, then those answers are shown in green, and can usually be inserted without any layout configuration necessary.
In some situations it may nevertheless be helpful to configure custom styling, e.g. to accommodate your house style.
ClauseBuddy's chatbot is aware of your current text selection. When you ask it to rewrite the selection, it will take only the paragraphs you selected.
General answers (i.e., those that you will not typically copy/paste into some document or email) will be shown with a light-purple background and can be copy-pasted through the small copy-button:
Answers that you will likely want to insert into your opened document (or email in Outlook) are shown with a light green background. They contain ClauseBuddy's special layout-aware insertion button:
The chatbot will frequently link back to relevant sections of the document. Just click on a paragraph to go to the corresponding paragraph in your document.
The chatbot has a legal filter built-in and will only respond to queries that relate to legal subject matters (in a wide sense). It will refuse to answer queries about completely unrelated subject matters, such as sports or pop songs. While there are always edge cases where the filter will be too strict or too relaxed, the filter is generally a good thing, as employers and law firms can assume that their subscriptions will not be abused for unrelated queries.
You can also upload a PDF-file or other Word-document, through the bottom-right corner. Just drag the document you want to "interrogate" on the upload area (or select it through your operating system's standard file section dialog box).
When you don't get the light-green background, be more explicit about the fact that you want the chatbot to write (or rewrite) something for you.
The chatbot will primarily focus on your current text selection. You may want to explicitly tell the chatbot that it should consider the other parts of the document if it doesn't get your subtle hints in this regard.
While you can upload a PDF-file and other Word-document, they are not as nice to interact with as your currently opened document. For example, you cannot click on it and scroll to the relevant section, as you can do within the currently opened document. That's probably OK for a quick question, but if you want to do a "deep" investigation, you may want to consider opening the Word-document in a separate window in Word and then open ClauseBuddy in that specific window.
In many cases you can actually ask a question in another language than the document's language.
The primary use-case of the chatbot is to quickly get a feeling about a document. Similar to how you would use other types of AI, please do not rely on the answers provided by the software without doublechecking yourself.
Try to be as explicit as possible in your wording. Longer and more detailed questions usually lead to better results, because the AI-filter will then perceive more semantic proximity between your question and relevant paragraphs.
Do not treat Generative AI engines as experienced lawyers with deep knowledge of specific legal subjects. Instead, think about them as very bright law students that have only one year of experience with legal subjects.
Be aware that your entire document, split into independent clauses, will be sent to Microsoft's semantic AI engine, for conversion to semantic datapoints. Microsoft provides strong confidentiality guarantees: your data will not be reused in any way, neither by Microsoft, nor by OpenAI.
Semantic search is ideal for situations where you want to quickly find something in long documents, and you are not sure which particular keyword is being used. After all, Word's standard function only works great when you know which keywords are used.
When you enter one or more keywords, ClauseBuddy will not only search for those exact keywords, but also for grammatical variations (e.g., "confidential" would also lead to "confidentially", and "disclose" would also lead to "disclosing" and "discloses") and for words that are — semantically speaking — closely related.
Examples of semantically related words:
"payment" will also find varations of the verb "pay", as well as adjectives such as "payable", nouns such as "cash" and "invoice" and "money"
"court" will also find variations such as "judge" and "trial" and "jury". Be aware that it may also find expressions such as "basketball court", if they would be present in your document
You can also search in a different language when you work on a multilingual server. For example, on the French/German/Dutch server, when entering a word such as "authorities" the software will automatically highlight the French/German/Dutch translations "autorités"/"Behörden"/"autoriteiten" (and semantically related words).
It then highlights all paragraphs that contain all those words. With a simple click, MS Word will then highlight that particular paragraph.
No GenAI involved. Unlike Chat with your document, which necessarily involves the use of a slow but intelligent Generative AI engine (such as GPT-4o), semantic search is completely calculated "locally" on ClauseBase's servers. Accordingly, no GPT-credits are being subtracted when using Semantic Search. This may be relevant for you from a compliance perspective.
The local calculation is also the reason why Semantic Search can be super quick — it can search through a document of 100 pages in less than a second.
Fast. DocChat can be relatively slow, because the GenAI-engine must read and understand your entire document. With each and every question you ask — even the tiniest follow-up question — it may very well be the case that the entire document will be resent for analysis. Semantic Search uses a lighter type of intelligence that doesn't go as "deep". It can therefore be much quicker than DocChat
Keyword-focused. Semantic Search is a replacement for MS Word's naive keyword-based search. DocChat is instead focused at deeper questions.
