Adaptive Intelligence

Adaptive Intelligence analyzes how your team uses Inscribe and recommends specific configuration changes so your workspace better reflects your risk policies. Each recommendation links to a pre-populated setup page in the Decision Engine, so you can review the logic and decide whether to apply it.

Recommendations are advisory. Nothing changes in your workspace until you review a recommendation and choose to set it up.

Recommendations are based on the ways your team engages with the Inscribe web app:

The more your team uses these feedback mechanisms, the more relevant your recommendations will be.

Adaptive Intelligence can suggest changes such as:

Reduce false positives

Your team regularly accepts documents that flag for a particular insight.

"For example, if font differences on bank statements are frequently accepted by reviewers, Adaptive Intelligence may recommend lowering that insight's severity."

Escalate signals that correlate with rejections

An insight appears more often on applications your team rejects.

"For example, if annotations on Wells Fargo bank statements appear more frequently in rejected applications, you may be recommended to add a high-severity override."

Catch outdated documents

Reviewers regularly reject documents that are too old.

"For example, if older bank statements are frequently rejected during review, you may be recommended to create an insight that identifies statements over 90 days old."

Block software linked to rejections

A specific software appears in the metadata of documents you reject.

"For example, if documents created with Quartz PDFContext are often associated with rejected applications, you may be recommended to add it to your software blocklist."

Step 1 - Receive your recommendations email

Workspace admins receive an email listing the configuration opportunities identified for your workspace. The email also summarizes your workspace activity over the last 30 days, including Customers added, User decisions, and Assistant queries. Each recommendation shows a short title and a one-line explanation of why it was suggested.

Step 2 - View your recommendations

Click View Recommendations in the email, or click any recommendation row, to open Adaptive Intelligence in the Decision Engine. You can also open it any time from the Decision Engine in the web app.

Step 3 - Review the recommendation

Click a recommendation on the Recommendations tab to open its configuration logic page. The logic is pre-populated, so you can see exactly what the recommendation would change before deciding. Recommendations are identified by the Adaptive Intelligence icon.

Step 4 - Set it up

If you want to apply it, click Set up. You can edit the pre-populated logic before saving. If a recommendation isn't right for your team, you don't need to take any action.

Backtesting is currently available for Custom Transaction Insights. If a recommendation creates one, use Estimate Impact on the configuration page to see how many customers from a sample would be flagged before you set it up. See our Custom Transaction Insights guide for details.

There are two common reasons.

Limited feedback activity. If your team hasn't used the feedback mechanisms much, there may not be enough data to generate recommendations. Encourage your team to accept / reject at the customer and document level, use 👍 / 👎 on Fraud Reasoning, and use the Assistant. Recommendations will follow as activity grows.

Your workspace is already well configured. If your team is engaging with these mechanisms and no recommendations appear, our analysis suggests no changes are needed right now. Adaptive Intelligence keeps monitoring your team's activity and will recommend changes if anything shifts.

Only users with Admin permissions receive Adaptive Intelligence emails and can set up recommendations. If you would like access, please reach out to your workspace Admins to change your permissions. Alternatively, ask your workspace Admins to review and apply recommendations on your behalf.

Recommendations are delivered by email to workspace admins and are available on the Recommendations tab of the Decision Engine. Once you set up a recommendation, the resulting insight, rule, override, or blocklist entry appears alongside your other configuration in the Decision Engine, and shows up in the Risk Analysis view when triggered.