-
Notifications
You must be signed in to change notification settings - Fork 416
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
77 changed files
with
896 additions
and
53 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,15 @@ | ||
| The Analytics page is designed to provide users with comprehensive insights into their AI Agents' performance. | ||
|
|
||
| It's your one-stop destination for understanding how your AI Agents are performing, engaging with users, and much more. | ||
|
|
||
|  | ||
|
|
||
| Conversations over time | ||
|
|
||
|  | ||
|
|
||
| Message rating: Good vs Bad evolution | ||
|
|
||
|  | ||
|
|
||
| Conversation origin worldmap |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,13 @@ | ||
| Sources are links to Datasources used to generate an answer. | ||
|
|
||
|  | ||
|
|
||
| Answer Sources | ||
|
|
||
| ### Disable Sources | ||
|
|
||
| You can disable this feature from your Agent global settings tab. | ||
|
|
||
|  | ||
|
|
||
| Disable Sources |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,10 @@ | ||
| If your Agent (AI chatbot) isn't producing accurate or truthful responses, consider the following steps to diagnose and address the issue: | ||
|
|
||
| <Note> | ||
| ℹ️ It's important to note that Large Language Models operate on a probabilistic basis, which means receiving the exact same response to a particular question is not guaranteed. | ||
|
|
||
| </Note> | ||
|
|
||
| - [ ] Are you able to reproduce the issue? | ||
| - [ ] Is this happening with new conversations also or only specific ones? | ||
| - [ ] Do you have the same issue when using more advanced models like gpt-4-turbo? |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,27 @@ | ||
| When your AI agent fails to accurately respond to a question, it is often due to the lack of explicit information or the absence of relevant data in the datastore it accesses. To enhance the accuracy of your agent or to uncover knowledge gaps, integrating a Q&A datasource is recommended. | ||
| This approach enables the agent to access explicit answers to specific questions. | ||
|
|
||
| Consider a scenario where an agent is queried, "What’s nuclear fusion?" Without information on nuclear fusion in its datastore, the agent might respond, "I’m sorry, I don't have the information you’re looking for” | ||
|
|
||
| <aside> | ||
| ℹ️ In this example, the knowledge restriction option is enabled; otherwise, the language model would have used knowledge from its training dataset. | ||
|
|
||
| </aside> | ||
|
|
||
|  | ||
|
|
||
| To rectify this, click the "improve" button (available only from the Inbox page). You will then be prompted to provide a correct answer for the question that stumped the agent. | ||
|
|
||
|  | ||
|
|
||
| This process automatically generates a new Q&A datasource within the linked datastore, which you can later edit or delete from the Datastore page. | ||
|
|
||
|  | ||
|
|
||
| Following this procedure ensures that the agent will correctly respond to similar questions in the future. | ||
|
|
||
|  | ||
|
|
||
| This method is effective across various languages, offering a versatile solution to improve your AI agent's performance 😎 | ||
|
|
||
|  |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,25 @@ | ||
| The Human Handoff feature significantly enhances customer interactions by providing a smooth transition from AI-driven conversations to direct human operator control. | ||
|
|
||
| This functionality ensures that users consistently receive accurate and empathetic responses at any point during their interaction. | ||
|
|
||
| You have the ability to intervene and take control from your AI Agent at any moment. | ||
|
|
||
| To activate Human Handoff mode, simply navigate to the Inbox page and click the "Reply" button. | ||
|
|
||
| **Note: When Human Handoff mode is activated, the AI Agent will be temporarily disabled.** | ||
|
|
||
|  | ||
|
|
||
| To return to automated responses and re-enable your AI, click the "Enable AI" button. | ||
|
|
||
|  | ||
|
|
||
| The Human Handoff feature seamlessly integrates with various communication platforms, including: | ||
|
|
||
| - Chat Bubble | ||
| - IFrame | ||
| - Standalone Page | ||
| - Crisp | ||
| - Slack | ||
|
|
||
| By leveraging these integrations, the feature ensures a versatile and efficient user experience across different messaging environments. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,6 @@ | ||
| Use message suggestions to display example messages the end user can pick from. | ||
| This is a great way to give your users hints about what type of questions they can ask to the Agent. | ||
|
|
||
|  | ||
|
|
||
| Message template demo |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,21 @@ | ||
| [Chaindesk](http://chaindesk.ai) is powered by large language models (LLM) like OpenAI’s ChatGPT. | ||
| As powerful as LLMs are, they can sometimes produce inaccurate or out-of-scope answers. This is called “hallucination”. | ||
|
|
||
| [Chaindesk](http://chaindesk.ai) provides multiple mechanisms/features to decrease hallucination. | ||
|
|
||
| ### Avoid duplicated content | ||
|
|
||
| Adding multiple times the same documents/information, even translated should be avoided. | ||
| Duplicated content can impact the Chatbot's performance. | ||
| It is best to add a piece of information only once, whatever the language | ||
|
|
||
| ### Use GPT-4 | ||
|
|
||
| Using GPT-4 can help to generate more accurate and truthful responses. | ||
| This model is also more respectful of the instructions given in a prompt. | ||
|
|
||
| The downside of GPT-4 is that it costs more credits and is slower than GPT-3.5-turbo. | ||
|
|
||
| ### Fine-tune your Agent’s answers | ||
|
|
||
| Use our latest feature, [Fine-tune Answers](https://www.notion.so/Fine-tune-Answers-90b51096e0474f579dd79f2f94e99f85?pvs=21), to correct an Agent's answer. |
Oops, something went wrong.