AI connectors

Bonitasoft now offers AI connectors that let you seamlessly integrate advanced language models like OpenAI and Mistral into your business processes. These connectors support three powerful use cases:

  • Generate text content from custom prompts

  • Classify documents into predefined categories

  • Extract structured data from unstructured text or images

By securely sending data from your processes to LLMs over HTTPS, you can enhance automation, improve decision-making, and streamline information handling—all without writing custom code.

The Bonita AI Connectors are available for Bonita 10.2 Community (2024.3) version and above.

AI Connectors flux

Getting started

To use a connector, add it as an extension dependency to your Bonita project. Choose the one related to your AI provider.

Currently supported providers:

  • OpenAI OpenAI

  • MistralAI Mistral AI

Image documents are not supported yet for Mistral connector due to a limitation of the underlying library we use.

Common configuration

Parameter name Required Description Default value

apiKey

false

The AI provider API key. Parameter is optional for testing purpose but obviously required with official OpenAI endpoint.

The connector will look for API key value in this order:

  • A system environment variable named AI_API_KEY

  • A JVM property named AI_API_KEY (-DAI_API_KEY=xxx)

  • the provided connector apiKey parameter

And at last it will use a dummy default changeMe value if no other value found.

changeMe

url

false

The OpenAI endpoint url. This parameter allows to use an alternate endpoint for tests.

Default to the official provider endpoint if not specified.

requestTimeout

false

The request timeout in milliseconds for OpenAI calls.

null

chatModelName

false

The model to use for chat like gpt-4o, gpt-4o-mini, llama, pixtral, …​ See respective documentation for detailed information about models.

  • OpenAI : gpt-4o

  • MistralAI : pixtral-12b-2409

modelTemperature

false

The temperature to use for the model. Higher values will result in more creative responses. Must be between 0 and 1.

Leave blank if the selected model does not support this parameter.

If the parameter is not present, the temperature will not be set in chat context.

null

Connectors

AI connectors have the capability to retrun structured data in JSON format. It is possible to pass a JSON schema to tell the LLM how to format response data.

When using a JSON schema, you must list in the required property, all the fields you want in the JSON response.

JSON schema sample
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "ProofOfAddress",
  "type": "object",
  "required": [
    "firstName",
    "lastName",
    "fullName",
    "fullAddress",
    "emissionDate",
    "issuerName",
    "identificationNumber"
  ],
  "properties": {
    "firstName": {
      "type": "string"
    },
    "lastName": {
      "type": "string"
    },
    "fullName": {
      "type": "string"
    },
    "fullAddress": {
      "type": "string"
    },
    "emissionDate": {
      "type": "string"
    },
    "issuerName": {
      "type": "string"
    },
    "identificationNumber": {
      "type": "string"
    }
  }
}

Ask connector

Take a user prompt and send it to OpenAI then return the AI response. The prompt text can ask question about a provided process document.

Parameter name Required Description Default value

systemPrompt

false

The system prompt to pass to the OpenAI endpoint. It helps to influence the behavior of the assistant and specify a default context. (ex: You are a personal finance advisor, providing guidance, etc.)

"You are a polite Assistant"

userPrompt

true

The user prompt content to send to the AI provider

sourceDocumentRef

false

The reference to the process document to load and add to the user prompt.

If not null, the connector will try to read the specified document and send it as an attachment to the user prompt. Supported formats are "doc", "docx", "pdf", …​ (see https://tika.apache.org/3.1.0/formats.html)

null

outputJsonSchema

false

The JSON schema that represent how to structure the JSON connector output.

null

The result can be a simple JSON object or one compliant with a provided JSON schema. This result will be placed as a map entry of type java.lang.String for the key named output.

Extract connector

This connector allow extracting information from a bonita document.

Parameter name Required Description Default value

sourceDocumentRef

true

The reference to the process document to load and add to the user prompt.

If not null, the connector will try to read the specified document and send it as an attachment to the user prompt. Supported formats are "doc", "docx", "pdf", …​ (see https://tika.apache.org/3.1.0/formats.html)

null

fieldsToExtract

false

The list of fields to extract from the given document.

The connector expect a list of String (like firstName, lastName, address).

null

outputJsonSchema

false

The JSON schema that represent how to structure the JSON connector output.

If a JSON schema is specified, the fieldsToExtract parameter is ignored.

null
You must provide at least one of fieldsToExtract or outputJsonSchema parameters.

The result can be a simple JSON object or one compliant with a provided JSON schema. This connector result will be placed as a map entry of type java.lang.String for the key named output.

Classify connector

This connector allows classifying a bonita process document according to a list of categories provided by the user.

Parameter name Required Description Default value

sourceDocumentRef

true

The reference to the process document to load and add to the user prompt.

If not null, the connector will try to read the specified document and to send it as an attachment to the user prompt. Supported formats are "doc", "docx", "pdf", …​ (see https://tika.apache.org/3.1.0/formats.html)

null

categories

true

The list of categories used to classify the given document.

The connector expect a list of String (like RIB, ID, …​). It is recommended to add a default category if none other matches such as Unknown

null

The result is a JSON String such as the following sample.

sample classification result
{
  "category": "xxx",
  "confidence": 0.9
}

The confidence score is defined as :

  • [0.0..0.3]: Very uncertain or guessing

  • [0.3..0.6]: Some uncertainty, potential ambiguity exists

  • [0.6..0.8]: Reasonably certain, minor doubt

  • [0.8..1.0]: Very certain, no doubt

This connector result will be placed as a map entry of type java.lang.String for the key named output.