8 OGL AI Assist (Generative Artificial Intelligence)

OGL AI Assist integrates Oracle Cloud Infrastructure Generative Artificial Intelligence (OCI Generative AI) capabilities into the OGL console to help content developers reduce time, add efficiency, and increase personalization of guide contents at scale. With the OGL AI Assistant ('AI Assist'), content developers can reduce the time needed to create content by utilizing AI to generate recommendations for tip text using a variety of prompts.

In this first iteration of the OGL AI Assistant, content developers can generate recommendations for tip text within Message guides only. Users can generate multiple variations of content recommendations, then incorporate the AI generated content into the message tool tip and customize, as needed. In future releases, the OGL AI Assistant will be extended to other guide types and console functionality.

Note:

This feature will be available to existing OGL customers in 24D.1 as an opt-in* feature and can be enabled upon request. If you wish to request this feature to be enabled on your OGL account, please submit a support ticket, or contact your OGL account manager.

Please include the specific APPIDs on which to enable the OGL AI Assistant.* Before opting-in to enable the OGL AI Assistant, please consult with your company or organization’s policies on usage of Generative AI functionality.

For all new OGL accounts (or APPIDs) provisioned after OGL 24D.1, the AI Assistant will be enabled by default. Customers who wish to opt-out and disable this feature on their accounts can submit a support ticket, or contact their OGL account manager.

Key considerations when using AI Assist:

1. Clients can start using AI Assist when looking for text content recommendations in their OGL tooltips ONLY for Messages. They can then add this content into their tooltips and amend as per usual (within the full editor).

   For example: “Write a service maintenance message for July 4th ”, or “Write a new release notes message”, or “Write a policy update notification.”

2. AI Assist captures language, tone, and text length.
3. Can regenerate the prompt if you're not satisfied with a given response.
4. With any AI prompt, it's best to be specific and to the point when writing instructions.
5. The message that is created uses the last generated response from AI (not earlier ones) – See below.

How to use the OGL AI Assistant for a Message guide:

1. From the My Content page, select the Create button to create a new content item.
2. In the New Content modal, select Message from the Type dropdown menu.

   This will automatically enable the OGL AI Assistant as shown below.

   

3. Enter a guide name for your message in the Display Name in Widget field and a Location URL, as required.
   
4. In How can I help you? field, enter a text prompt for what you would like the OGL AI Assistant to generate: (For example, “Create a system downtime alert message for this Sunday from 5:00AM - 7:00AM”) Select the Send button (or press Enter) to generate a response.
   

5. The OGL AI Assistant responds by generating a recommendation for the message text based on your prompt.
   

   Below are the different response buttons that allow you to:

   • Regenerate a new response.

   • Copy the response text to your clipboard.

   • Clear the response and start over with a new prompt.

   • Create to create a new Message guide using the response as the tip text.

   6. If you’re not satisfied with the recommended response, you can ask the AI Assistant to generate another response by selecting the Regenerate button. Each time you regenerate a new response, you’ll be able to toggle between the responses by using the Previous and Next Response buttons.

   

Note:

Selecting the Create button at the bottom of the modal will automatically paste the last generated text response. Even if you navigate to a previous response, such as < 2/3 >, selecting Create will paste the text from < 3/3 > into a new Message guide.

To copy the text from a previous response, use the Copy button to copy it to your clipboard and paste it into the tooltip in the guide editor or into a text file for later use.

Tips for using the OGL AI Assistant

• Refine your prompts
  • If the AI Assistant response is too long, you can refine it by adding additional prompts within in the same session. For example, “Make it shorter and more formal.”
    

• Types of prompting questions
  • You can prompt the AI Assistant by adding in verbiage around language translations (e.g., “Write this message in Japanese”), tone (e.g., Corporate or Casual), and character/word length (e.g., “in two

    sentences” or “Write a brief message…”)

  • You can provide the context of the audience you’re writing for (e.g. For HR Managers)
• Prompting Guidelines
  • Best guideline for the prompting is to be specific with your prompt in what you want, how it should be presented and who it is for. You can use Do and Don’t commands.
  • Some articles that may help you include: https://huit.harvard.edu/news/ai-prompts & https://news.microsoft.com/source/features/ai/the-art-of-the-prompt-how-to-get-the-best-outof-generative-ai/
• Guidelines for invoking prompt into message
  • Note 1 – when refreshing your AI prompt the last copy of the prompt will be copied into the editor experience as designed. To use an earlier version, please use the copy button and instead paste this inside the editor.
  • Note 2 – At this stage in time, AI Assist can only use this for generating message text. At a future state we will be rolling out the capability to build guides and summarizing analytics reports

Known Issues/Expected Behavior for OGL AI Assistant:

• In this current release, the scope of functionality for the OGL AI Assistant is intended to generate recommendations for tip text in a Message guide only. This feature will be further expanded to other guide types and other application functionality in future releases.
• Expected behavior – Auto-Generated page condition will be created upon Save & Exit.

  Condition: Display this guide in Autoload When Page has URL Matching {Location URL, whether manually entered or default Location URL}

• Expected behavior - Selecting the Create button at the bottom of the modal will automatically paste the last generated text response. Even if you navigate to a previous response, such as < 2/3 >, selecting Create will paste the text from < 3/3 > into a new Message guide. To copy the text from a previous response, navigate to the response you wish to copy, select the Copy button to copy it to your clipboard, and paste it into the tooltip in the guide editor or into a text file for later use.

Data Handling in Generative AI

Learn how OCI Generative AI handles user data.

• Does OCI Generative AI retain customer-provided training data used to fine-tune a custom model?

  No. The customer stores and manages their training data in their own customer tenancy (commonly inside an OCI Object Storage bucket). OCI Generative AI's fine-tuning job uses this training data to train a custom model for customer. OCI Generative AI doesn't retain this training data beyond the duration of this training job. This training data is solely used for building a custom model for this customer. The custom model is also a resource that's managed by the customer. This training data is not used to improve the general use cases for OCI Generative AI.

• Does OCI Generative AI retain customer-provided prompts and inputs used for inferencing, on the large language models (LLMs)?

  No, OCI Generative AI doesn't retain customer inputs. A user's input on an inference call is sent to the LLM and a response is generated by the LLM that's returned to the user. Both input and output are not stored inside OCI Generative AI.

• Does OCI Generative AI share prompts and responses, fine-tuning training data, or fine-tuned custom models with third-party model providers such as Cohere or Meta?

  No.

• Is the training data encrypted?

  The data is encrypted both at rest and in transit. The training data is deleted from the fine-tuning cluster as soon as the training job is completed.

  Training data for fine-tuning a model is always double encrypted, by default, by Oracle-managed AES-256 encryption and optionally by customer-managed private keys through OCI Vault service. Customers can delete the data at any time. Oracle encrypts all the data in motion with TLS 1.2.