Exploratory module on the use of AI in Simplicite
To import this module, you have two options:
- Create a module named
AIBySimplicite
. - Set the settings as:
{
"origin": {
"uri": "https://github.com/simplicitesoftware/module_ai.git"
},
"type": "git",
"branch": "<6 or 6.1>"
}
- Click on the Import module button.
- Navigate to the App Store in the Project domain.
- Go to the Tools tab.
- Select "AI integrations" from the list.
- Follow the on-screen instructions to import the module.
To configure the AI connection, utilize the configuring the wizard
process in the AI assistant
domain in extended mode.
Select the model (LLM) you wish to use, configure the API key, and the URL of your API:
Follow the process:
After configuring, you will obtain the details of your now active configuration. This summary is available on the homepage of the domain.
Parameters details
Defines the number of messages in the history used for the context. The greater this number, the more tokens the request uses, but the more relevant is the response.
optional
Limits the number of tokens in the AI response.
Maximum number of tokens in a usual context.
Maximum number of tokens for technical calls.
- AI domain
- LLM and API configuration
- AI-assisted module creation or completion process
- Data generation process
- Chatbot for Business with context
- List of LLMs currently configurable.
Use AI to co-create or update a module based on your business needs
Module generation
in AI assistant
domain.
Use AI to generate test data for a module.
Data generation
in AI assistant
domain.
A shortcut to a chatbot contextualized according to the form on which it is opened.
The user must have AI_BUSINESS
rights.
The Personal data
, Confidential data
and Intimate
fields are not sent to the AI.
The chatbot can be contextualized to the current user's scope, including knowledge of the UI and business rules. To achieve you have to pregenerate the ui and rules description:
- Open the settings of the scope to which you want to add a context (in the User interface => views => Home page submenu).
- Go to the tab of the link with
Group gui description
. - Create a new
Group gui description
by choosing the desired language. - You can refine the description by selecting the groups to be considered for this description (by default, the groups of rights in the scope are taken into account).
- The chatbot open in this scope will have UI context
- The lang (to get context in the user language)
- Auto generated Description of the ui and the rules
- The descritpion will contaigne all visible domaine and object for the listed groupe with GUIDescription has yes
Important: remember to regenerate this description when there are changes to the rights and rules of your objects
Generating AI graphs on a module's data
In a view
add an external page
of source External object
: AIMetricsChat?module=<you_module_name>
For detailed examples on how to generate modules, interact with business chatbots, and visualize metrics, please refer to the example available at the following link: Explore AI Assistant Prompt Examples
Name | Type | Required | Updatable | Personal | Description |
---|---|---|---|---|---|
aiPrvProvider |
char(100) | yes* | yes | - | |
aiPrvDefaultUrl |
url(100) | yes | yes | - | |
aiPrvDataModel |
text(1000) | yes | yes | - | |
aiPrvModelsUrl |
url(100) | yes | yes | - | |
aiPrvHelp |
text(5000) | yes | - | ||
aiPrvCompletionUrl |
url(100) | yes | yes | - | |
aiPrvPingUrl |
url(100) | yes | - |
EndHome
:Begin
:End
:IsParam
:NoParam
:SelectModule
: Selection moduleConfirm
:GenData
: AI call for data gen
Auomatic model generation process
Begin
:End
:EndHome
:Choice
:CreateModule
:TranslateDomain
:NewScope
:GrantUser
:isParam
:NoParam
:SelectModule
: Selection moduleSelectGroup
: Selection groupSelectDomain
: Selection Domaininteraction
: Chat with the AIPrompt
: Direct chat with the AIAI
: AI for modelingGeneration
: Module generation based on ia feedbackRemoveModule
: Remove module
isGlobalParam
:Global
:SelectProvider
:Auth
:OtherParam
:Begin
:End
:
Chat interface
Front tool
AI contextual chat for personalised metrics. Use parameter "module" to specify context.
Process resource
Tool
Local API
Chat in shortcut