What's New at spektr

Features

Arithmetic and named formulas in transaction rule conditions

You can now combine numbers, customer record fields and aggregation values with +, −, ×, ÷ and brackets directly in a transaction rule condition. A new Formulas section in the transaction rule dialog lets you create named formulas and reuse them across conditions, including as the base of "% of" and "× of" comparisons. A division by zero or an empty value makes the condition not met. Per-customer thresholds, for example volume × (1 + tolerance / 100), can now live in the rule itself instead of being pre-computed outside spektr.

What this changes for you: Rules that relied on values pre-computed outside spektr can now carry the math in the condition. Existing rules are unchanged.

One row per rule in test run reports

The Test Version dialog now offers a "Rule test run report" alongside the existing backtest, which is renamed "Client test run report". The new report writes one CSV row per customer, per simulated date, per selected rule, with a Rule output column showing whether the rule fired. You pick the rules to include with checkboxes, none selected by default, and only the aggregations a rule actually uses carry values.

What this changes for you: If you run backtests, the dialog now shows two report types. Existing reports keep working under the new name.

Cases keep the numbers a rule fired on

When a transaction monitoring rule fires, the case now stores the measured values behind each condition. The trigger card shows every condition with its measured value, threshold and a Met or Not met badge, and the AI case summary quotes the same numbers, so it can no longer cite a figure the rule did not fire on. If the rule fires again on an open case, the card shows the latest numbers. This applies to cases created by scheduled and batch processes.

What this changes for you: Case reviews now show the actual figures behind each condition. No setup change needed.

D&B ownership monitoring compares like-for-like

Rolling out behind a feature flag. D&B monitoring now compares ownership percentages like-for-like: the direct holding recorded at onboarding is compared with D&B's direct figure, and deeper owners are tracked on D&B's indirect percentage. Alerts state which measure changed, the old and new value, the owner's degree, and whether the change crosses the 25% UBO threshold. Confirming a change on a direct owner updates that link and recalculates totals and UBO verdicts. Confirming a change on a deeper owner no longer overwrites your ownership structure.

What this changes for you: No change until the flag is enabled for your workspace. Once it is, ownership-change alerts fire only on genuine changes and state exactly which measure moved.

Improvements

Clearer placeholders for missing report values

Missing-value placeholders on case reports now say whether a field is filled by the system or needs an analyst, replacing the previous generic placeholder text. Editable and read-only fields share one grey background, and amber highlighting now marks only the gaps an analyst needs to fill.

What this changes for you: Reports look different where fields are empty. Worth a heads-up to anyone who completes reports.

Paused processes visible in the Ongoing processes list

Ongoing processes paused by a Multiple KYB Hits alert now appear in the Ongoing processes list, where they were previously hard to find.

What this changes for you: Check the Ongoing processes list for paused work that was previously hidden.

This release also includes bug fixes, platform and performance work.

‍