> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getpg.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent

> Conversational AI that knows your accounts  -  ask questions and get answers grounded in real intelligence data

A conversational AI that knows your accounts. Ask questions about any company and get answers grounded in real intelligence data, not generic web search.

When you open a conversation, the Agent already has access to the full Intelligence profile: strategic priorities, contacts, tech stack, financials, hiring data, enrichment history, and previous conversations. Every answer is grounded in what PG:AI knows about the account.

{/* IMAGE PLACEHOLDER: Screenshot of Agent main view */}

## Key capabilities

* **Conversational research** with full account context
* **Real-time web search** and deep research on demand
* **Source citations** for every claim
* **Conversation memory**  -  previous research and follow-ups are saved
* **Content streaming** for real-time responses
* **Multiple conversation threads** per account

## Who it's for

| Role               | How they use it                                                                             |
| ------------------ | ------------------------------------------------------------------------------------------- |
| Account executives | Preparing for meetings, investigating accounts, getting quick answers to specific questions |
| SDRs               | Researching accounts for outreach                                                           |
| Sales leaders      | Getting quick account summaries                                                             |

## Effective question patterns

The Agent draws from different account intelligence modules depending on what you ask. Here are the most effective patterns.

### Strategic questions

* "What are \[Company]'s top strategic priorities for this year?"
* "What challenges is \[Company] facing in their market?"
* "How does \[Company]'s strategy compare to \[Competitor]?"
* "What are the main risks to \[Company]'s business?"

<Info>
  These draw primarily from the **Strategic Insights** module: priorities, goals, SWOT, and division intelligence.
</Info>

### People questions

* "Who leads the platform engineering function?"
* "Who reports to the CTO?"
* "Find me the VP of Sales or equivalent."
* "What's the career history of \[Contact Name]?"

<Info>
  These draw from **Contacts & Org Chart** data.
</Info>

### Technology questions

* "What cloud technologies are they using?"
* "How does their tech stack compare to \[Competitor]?"
* "What technologies are trending in their recent job postings?"
* "Do they use \[specific technology]?"

<Info>
  These draw from **Tech Stack Intelligence** and **Jobs** data.
</Info>

### Hiring and investment questions

* "What roles are they hiring for right now?"
* "How has their hiring changed in the last quarter?"
* "Are they investing in \[specific area]?"
* "What does their hiring pattern signal about their priorities?"

<Info>
  These draw from **Jobs & Hiring Signals** data.
</Info>

### Financial questions

* "What's their current revenue and growth rate?"
* "How do their margins compare to competitors?"
* "What did analysts say in their last earnings call?"
* "Are they under financial pressure?"

<Info>
  These draw from **Financial Intelligence** data.
</Info>

### Meeting preparation

* "Prepare a meeting brief for my call with \[Company] tomorrow."
* "What are 5 discovery questions I should ask based on their strategic priorities?"
* "What's the strongest value alignment between our solution and their goals?"

<Info>
  These synthesise across multiple data modules to produce comprehensive output.
</Info>

## Tips for better responses

<Steps>
  <Step title="Be specific">
    "What cloud technologies are they using?" is better than "tell me about their technology."
  </Step>

  <Step title="Reference the account">
    The agent knows which company you're viewing. You don't need to restate the company name in every message.
  </Step>

  <Step title="Ask follow-ups">
    The agent remembers the conversation. "Go deeper on that" or "How does that compare to last year?" works naturally.
  </Step>

  <Step title="Request research">
    If the Intelligence profile doesn't have what you need, ask the agent to research it. "Research their recent partnerships" triggers a web search.
  </Step>

  <Step title="Ask for formatting">
    "Summarise this as bullet points" or "Format this as an email" works within the conversation.
  </Step>
</Steps>

## Limitations

<Warning>
  * The agent's knowledge of an account is only as good as the Intelligence enrichment. Accounts with minimal enrichment produce thinner responses.
  * Real-time web search may not surface very recent events (last few hours).
  * The agent does not have access to your CRM data, email, or internal documents unless they've been uploaded to PG:AI.
</Warning>

## Related modules

<CardGroup cols={2}>
  <Card title="Canvas" icon="file-lines" href="/studio/canvas">
    Turn research conversations into polished documents  -  executive briefings, engagement plans, and more.
  </Card>

  <Card title="Workflows" icon="diagram-project" href="/studio/workflows">
    Automate multi-step research and content generation across accounts at scale.
  </Card>

  <Card title="StudioMail" icon="envelope" href="/studio/agentmail">
    Give the Agent access to email context for better follow-ups and responses.
  </Card>
</CardGroup>