Overview

Google AI provides access to Gemini models for generating high‑quality text. Persona’s integration with Google AI lets teams automate content creation and case assistance directly from Workflows and Cases—so you can summarize evidence, draft outreach, and standardize internal notes without leaving Persona.

This integration helps fraud, risk, and compliance operations reduce manual writing and improve consistency by invoking Gemini to produce context‑aware text using data from Persona.

Benefits

Faster Case Summaries: Automatically generate concise, standardized summaries for reviews and audits.

Consistent Outreach: Draft customer communications that reflect your policies and tone, reducing manual edits.

Agent Copilot: Suggest next steps, internal notes, and decision rationales based on available case details.

Integration Features

Persona’s Google AI integration focuses on safe, configurable content generation that can be embedded in your existing review flows. Key capabilities include:

• Text Generation: Produce tailored narratives, explanations, and messages from case context and prompts.
• Configurable Outputs: Control variability and length with parameters like temperature and max tokens for predictable results.
• Multi‑turn Prompts: Provide conversation history to keep responses consistent across steps in a workflow.
• Operational Reporting: Access token usage metadata returned by Google AI to understand cost and response size at a glance.

Setting up the Google AI integration

Prerequisites

To set up the Google AI integration, ensure you have:

• Admin access to your Google AI account
• Necessary API permissions to access Google AI credentials

Setting up the Google AI Credentials

1. In the Persona Dashboard, navigate to Integrations > Marketplace and click on Google AI.
2. Click on Add Credential at the top of the page.
3. Enter a Nickname for Credential and click Continue.
4. Paste your Google AI API key from Google AI Studio and save the credential.

Tip: You can rotate credentials anytime by adding a new credential and updating your workflows to reference it.

Using the Google AI integration in a workflow

1. Create a new workflow, or open an existing workflow you’d like to update.
2. Add a new Action step > Integrations.
3. Select the Google AI integration and choose your Google AI credentials.
4. Configure the input fields (prompt, optional conversation history, and generation settings).
5. Save and publish the workflow.

Google AI Operations Overview

In addition to syncing field values, Persona can generate summaries, draft messages, and create structured internal notes using Google AI’s API. These actions support seamless two-way workflows, letting teams manage investigations without switching platforms. See below for a comprehensive list of available Workflow Action steps and possible configurations for the Google AI integration:

Generate Content

Generates a model response based on your prompt and optional conversation history. Use this step to create case summaries, draft customer messages, or propose internal notes aligned to your review guidelines. Responses can be tuned for tone, determinism, and length to meet operational needs.

Configuration Steps:

• Provide values for required fields:
  • Model Variant: Choose a supported Gemini model for text generation (see Google’s model documentation for options).
  • Content: Enter the prompt or supply a sequence of content entries that represent the conversation so far.
• Optionally add:
  • Conversation History: Provide multiple content entries with roles (e.g., user/model) to maintain context across turns.
  • Temperature: Adjust randomness; lower values yield more deterministic outputs.
  • Max Output Tokens: Cap response length to control verbosity and cost.
  • Top P and Top K: Fine‑tune sampling behavior for quality vs. diversity.
  • Stop Sequences: Specify strings where the model should stop generating.
• Output:
  • Candidate Text: The generated response content.
  • Usage Metadata: Token counts for prompts and responses to inform monitoring and cost control.

Notes:

• Input must include at least one content entry with text.
• Model availability and capabilities vary; review Google AI’s model documentation to select the right variant for your use case.

FAQs

What permissions are required in Google AI?

You need access to create and manage API keys and permission to use the Gemini API with your organization’s billing and policy settings.

Which models can I use?

You can use supported Gemini text models shown in Google AI’s model documentation. Choose a variant that balances latency, cost, and capability for your workflow.

Can I control output length and style?

Yes. Use generation settings such as Temperature, Max Output Tokens, Top P, Top K, and Stop Sequences to guide response tone, determinism, and length.

Does this integration support images, audio, or streaming?

This integration focuses on generating text from text inputs. File inputs and streaming responses are not supported through this integration at this time.

Where do I see token usage?

The Generate Content step returns usage metadata (e.g., prompt and response token counts) that can be surfaced in Workflow runs and logs for monitoring.