Spark Basics

Q: Why Spark vs. ChatGPT, Claude, Bing Chat, [insert favorite GPT product here]?

Spark has a few key advantages over these other tools, but like everything, there are tradeoffs

  • Spark gives you unlimited access to the latest OpenAI model for free.
  • Using Spark ensures your data and prompts are NOT submitted into the LLM to potentially surface in other responses
  • Spark is being built to meet the specific needs accounting professionals have within their firms in order to serve their clients. Your feedback helps ensure we’re meeting the needs of users like you.
Q: What model does Spark use?

Spark uses the most recent Large Language Model and features available from OpenAI via their API. The current version being used is GPT-4 Turbo. We make every effort to keep this updated as OpenAI makes new models available.

Q: How updated is the information the model includes in responses?

The current “knowledge cutoff” is April 2023. Any events, rules or regulations that have been updated since then will not be included in responses.

Q: Does Spark have access to the broader internet for more updated information?

Web search, and the ability to ensure you’re getting the most up-to-date information is not currently available within Spark, but it is on our roadmap and a feature we intend to provide for our users in the future.

Q: How much does Spark cost?

Spark is currently free for all users and includes all available features within the product.

Spark Features

Q: When I copy/paste an answer from Spark into another application, the formatting doesn’t translate correctly. What can I do?

Unfortunately, each application interprets formatting differently. Getting the format right may take some trial and error. Here are some things you can try:

  • Copy from Spark using the clipboard icon
  • Copy from Spark by selecting all the text and copying.
  • In your destination app, Edit>paste and match formatting or Edit>paste special… Formatted Text (RTF)
  • In Word, specifically, depending on the formatting within Spark, you may find that it put borders/boxes around your content. To remove, select all your content, go to “borders and shading”, click “none” and select “Paragraph” in the “apply to:” drop down. Repeat and select “text” in the “Apply to:” dropdown
Q: What is the content library used for?

The content library can be used for a variety of use-cases:

  • Upload your employment documentation (Benefits information, Employee handbook, etc.)
  • Add collections of documents for a specific project (legal cases or other documents needed to advise a client)
  • Add a collection for each client to add notes and transcripts
  • Add articles, podcast transcripts or other resources from your favorite content creators

Once you have a collection, you can have a conversation with that specific set of documents to retrieve information only found within those documents.

Q: Can I reuse prompts so I don’t have to copy/paste from past prompts or other documents I’ve created?

When you have a prompt setup the way you like it, you can click the [Save] button located to the right of the prompt in your chat. You can save it to your saved prompts by giving it a name. When you’re ready to use it again in the future, you can simply click “Browse Prompts” at the bottom of the chat window and locate it within “My Prompts” to start a new chat with your prompt.

Security & Privacy

Q: Will my data be used to train the model?

A: Your data and any prompts/questions submitted will not be used by Spark, or by OpenAI, to train the LLM.

The Spark team uses anonymized data to better understand how users are using the product so that we can improve the experience and enhance our features and capabilities.

For more information regarding OpenAI security visit https://openai.com/security.

Q: Who within my team can see what?

We are continuing to build out team sharing features and functionality, but you can be confident that no one in your team can see anything unless you explicitly share it with them - Saved prompts, assistants or content from a content library.

Q: How does Spark protect the privacy of data I enter into a prompt?

A: Your data is securely protected on our servers, and because we use OpenAI’s API, your data is not incorporated into their LLM. However, we recommend limiting input of PII (Personally identifiable information) and client information as a best practice.