> For the complete documentation index, see [llms.txt](https://docs.compliance.phalcon.blocksec.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.compliance.phalcon.blocksec.com/manual/screening/ai-risk-interpretation.md).

# AI Risk Interpretation

Lumi helps you understand an address screening result, identify findings to review first, and check the evidence behind them. You can ask about the overall result or focus on a specific risk category.

AI risk interpretation provides supporting analysis for human review. It does not change the displayed risk level or make a final compliance decision.

## Start an Interpretation

1. Open a screened address from the **Address List**.
2. Click **Ask AI** beside a risk section.
3. Check the address and **Screened at** time shown in the Lumi panel.
4. Read the interpretation, then select a suggested question or enter your own follow-up.

You can also open Lumi and ask it to interpret the current address screening result. Opening Lumi alone does not start an interpretation or use your AI interpretation allowance.

If the address has not been screened, screen it first. Lumi does not initiate screening automatically.

<figure><img src="https://1419887533-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FPfx6yOSRmp0GCnHv5rhy%2Fuploads%2Fgit-blob-b192361b8e0a45b0b6086e9475107d7d2037e615%2Fai-risk-interpretation-entry.png?alt=media" alt="Address details page showing Ask AI beside risk sections and the Lumi panel with the selected address, screening time, and risk interpretation"><figcaption><p>Use Ask AI to interpret a risk section, then review the answer and supporting evidence in Lumi.</p></figcaption></figure>

### Choose a Focus

| Section              | What Lumi helps you understand                                                                                                             |
| -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ |
| **Risk Summary**     | Which findings deserve attention first and what evidence supports that review priority.                                                    |
| **Entity Risk**      | Risk indicators associated with the address itself, including the available label or list information.                                     |
| **Behavioral Risk**  | Detected patterns, such as large transfers, high-frequency transfers, or transit address behavior, and the available metrics.              |
| **Interaction Risk** | Exposure to risk sources, including incoming or outgoing direction, direct or indirect exposure, amounts, percentages, and trigger status. |

## Read the Answer

An interpretation highlights the most relevant findings and explains what you can verify next. Depending on the available evidence, it may include:

* **Key findings** — The screening facts relevant to your question.
* **Evidence** — Supporting indicators, amounts, directions, behavior metrics, or linked alerts.
* **Review guidance** — Information to check before reaching your own conclusion.
* **Limitations** — Missing details or restrictions that affect the answer.
* **Suggested questions** — Follow-up questions based on the same screening result.

A highlighted evidence card may accompany the answer. It provides a reference to a finding rather than a separate risk assessment.

### Ask a Follow-Up or Check the Evidence

Suggested question buttons send a follow-up question to Lumi. They continue the conversation using the same screening result.

Evidence navigation buttons take you to the supporting information in the product. Depending on the available evidence and your access, a button may locate a risk section, highlight an indicator, or open a related alert or rule.

Checking evidence does not initiate a new screening. Navigation options appear only when the supporting information can be located.

## Understand the Data Scope

Each conversation uses one screening result, identified by its address and **Screened at** time. Follow-up questions use the same result; switching pages or re-screening does not update the conversation automatically.

To discuss the latest result, start a new interpretation. Evidence links for earlier results may be unavailable. Lumi flags missing data rather than silently using a newer result.

{% hint style="info" %}
AI interpretations can contain errors. Verify the supporting evidence before making a decision. Lumi does not resolve alerts, change Risk Engine settings, initiate screening, block or freeze funds, or create or submit SAR/STR reports through this feature.
{% endhint %}

## AI Interpretation Allowance

The following allowances apply during the AI risk interpretation Launch Offer, which runs through **December 30, 2026**.

The AI interpretation allowance is counted by **screening result**, not by question. It is separate from your screening quota.

### Free and Credits Plans

Free and Credits users have the same AI interpretation allowance:

* You can unlock AI interpretation for **3 different screening results over the lifetime of your account**. This allowance does not reset monthly.
* A result is unlocked only after its first interpretation is successfully returned.
* Follow-up questions and interpretations from other modules of the same unlocked result do not use another allowance.
* Each unlocked result supports unlimited follow-up questions.
* Re-screening an address creates a new screening result. Interpreting that new result requires another unlock.

Opening Lumi, a failed or cancelled request, an unconfirmed target, or an access or allowance rejection does not use an allowance.

After all 3 results have been unlocked, you can continue discussing those results. To interpret an additional result, follow the upgrade prompt.

### Paid Subscriptions and Enterprise Trial

Users with an active paid subscription or Enterprise Trial can unlock an unlimited number of screening results and continue asking about them during the offer. Credits users have the same allowance as Free users, as described above.

AI interpretation respects your existing access permissions. It does not reveal screening data or rule details that your account cannot access.

## If an Interpretation Is Unavailable

* **The request failed:** Select **Retry**. The retry uses the same question and screening result. A failed request does not use an allowance.
* **You do not have access:** Contact your administrator to request access to the screening result.
* **Your Free or Credits allowance is exhausted:** Continue using an unlocked result or follow the upgrade prompt to interpret a new one.
* **The evidence is incomplete:** Review the stated limitation and check the available product evidence.
* **The conversation refers to an earlier result:** Check its screening time and start a new interpretation if you want to discuss the latest result.
