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

# Reporting & Treasury Insights Agent

> On-demand and scheduled cross-chain operational intelligence from a read-only agent — every figure computed by WalletSuite, selected and explained by the model.

The Reporting & Treasury Insights Agent is one of WalletSuite's reusable agent templates — a concrete starting point you configure and embed rather than build. It turns your wallet and transaction data into **on-demand and scheduled cross-chain operational intelligence**: balances, trends, exceptions, and treasury exposure, on demand or on a fixed cadence.

It is not a generic chatbot. It runs at the [R0 execution level](/ai-agents/execution-levels), so it can read and explain but can never prepare, sign, or broadcast.

<Info>
  **Read-only by construction.** This template is bound to the **R0** read-only tool profile. Its credential carries no signing scope, so it never reaches the authorizer or signing path — see [Agents & Authority](/core-concepts/agents-and-authority). To move value, use a higher execution level; see [Execution Levels](/ai-agents/execution-levels).
</Info>

## How a report is produced

Three roles, and none of them let the model compute a figure or touch value:

<Steps>
  <Step title="A deterministic scheduler starts the report">
    Reports run on demand or on a fixed cadence.
  </Step>

  <Step title="WalletSuite services compute and aggregate every figure">
    Balances, volumes, fees, settlement status, exposure, and concentration are computed and aggregated by WalletSuite services behind the scenes.
  </Step>

  <Step title="The agent selects the relevant findings and explains them">
    The model's only job is to choose which findings matter and narrate them in plain language. It produces no numbers of its own.
  </Step>
</Steps>

## What every report carries

Every report is self-describing and traceable back to source:

| Field               | Meaning                                                                  |
| ------------------- | ------------------------------------------------------------------------ |
| Data window         | The time range the report covers.                                        |
| Freshness timestamp | When the underlying figures were last computed.                          |
| Template version    | The report template that produced it.                                    |
| Record references   | Pointers to the underlying transactions, balances, and policy decisions. |

Because every figure references the records it came from, a report is verifiable end to end — see [Audit Trail](/security/audit-trail).

## Supported reports

<CardGroup cols={2}>
  <Card title="Cross-chain balances & allocation" icon="scale-balanced">
    Native and token balances across every chain in the wallet grant, with allocation breakdowns.
  </Card>

  <Card title="Volume, fee & settlement trends" icon="chart-line">
    Movement, cost, and settlement patterns over the report window.
  </Card>

  <Card title="Pending, failed & delayed transactions" icon="triangle-exclamation">
    Summaries of transactions that are still pending, failed, or delayed against expected settlement.
  </Card>

  <Card title="Wallet health & operational exceptions" icon="heart-pulse">
    Operational anomalies and exceptions surfaced across your wallets.
  </Card>

  <Card title="Treasury exposure & concentration" icon="layer-group">
    Asset and counterparty exposure with concentration analysis.
  </Card>

  <Card title="Scheduled daily / weekly / month-end" icon="calendar-days">
    The same report set delivered on a daily, weekly, or month-end cadence.
  </Card>
</CardGroup>

## Configure and embed

<Steps>
  <Step title="Provision a hosted agent credential">
    Create a scoped, revocable agent key pinned to the **R0** read-only tool profile. This is the agent's identity.
  </Step>

  <Step title="Authorize the agent">
    Your server authorizes the agent under the read-only tool profile; the raw key stays server-side and only a short-lived token reaches the agent. See [Connect your agent](/ai-agents/token-exchange).
  </Step>

  <Step title="Deliver on demand or on schedule">
    Request a report on demand, or register a daily, weekly, or month-end cadence.
  </Step>
</Steps>

The read tools this template can call are listed in the [Tool Reference](/ai-agents/tool-reference). If you are deciding how to embed the agent, start from [Choose Your Surface](/ai-agents/choose-your-setup).

## Reporting agent vs. a Deterministic Workflow

Both compute every figure deterministically inside WalletSuite. The difference is whether a model is in the loop at all:

* **Reporting agent** — the model selects and explains findings within the R0 read-only profile. Use it when you want narrative operational intelligence, not just raw numbers.
* **Deterministic Workflow** — a fixed pipeline with no model in the loop. Use it when you want the same figures with zero interpretation.

See [Choose Your Surface](/ai-agents/choose-your-setup) to pick between them.

## Next steps

* Read [Execution Levels](/ai-agents/execution-levels) to understand R0 and how higher levels unlock write actions
* Read [Connect your agent](/ai-agents/token-exchange) for how the agent authorizes and connects
* Read [Choose Your Surface](/ai-agents/choose-your-setup) to configure and embed
* Return to the [AI Agents Overview](/ai-agents/overview) for the full agent model