Skip to content
Introducing complead. One AI-native platform for financial crime compliance. Read the story
New Ready Integrations available Check the new integrations

Nova, Agentic AI and the Widget Builder: complead June 2026 updates

In short

The June release gives the team AI colleagues. Nova, an AML assistant, answers questions about your data in a chat. Agentic AI acts on cases in the background, following rules you draw on a canvas. AI Suggest reaches transaction monitoring, KYC forms start from templates and screen applicants on submission, and the Widget Builder lets you design your own dashboard.

A compliance team spends much of its time on two kinds of work. One is looking things up: how many cases are open for a segment, which customers moved to high risk this week, what the history of a profile says. The other is applying decisions the team has already made many times, such as closing a namesake alert that clearly doesn't match, or sending a certain kind of case to a senior analyst.

Both are necessary, and both take time away from the cases that need real judgement. Last month AI Suggest started proposing decisions for name screening. In June we took the next steps, with an assistant to ask and agents that act on the rules you set. The highlights:

  • Nova, an AML assistant, answers questions and analyses your data in a chat, in beta.
  • Agentic AI handles cases in the background, following rules built on a drag-and-drop canvas.
  • AI Suggest reaches transaction monitoring alerts.
  • KYC forms start from ready-made templates and can screen the applicant as soon as the form is submitted.
  • The Widget Builder lets you design widgets from your own data and add them to your dashboard.

Nova, the AML assistant

Ask questions about your compliance data

The problem. Simple questions about the state of the work often needed a report, a filter combination or a colleague who knew where to look.

What changed. Nova, our AML assistant, launched in beta.

How does Nova work?

Nova opens from a tab at the side of the screen, so it is available wherever you are in the platform. You type a question in plain language, and Nova reads the relevant data in your organization to answer it. You can ask it to summarize the open work, to look into a customer or to explain what stands out in a set of cases, and continue the conversation with follow-up questions.

A few questions teams start with, as illustrations: "How many name screening cases are open and unassigned?", "Which customers had their risk level raised this week?" or "Summarize the history of this customer before I review the alert." Each answer is a starting point for the review, not a replacement for it.

Nova works within the permissions of the user who asks, and its answers stay inside your organization's data. As with other AI features, it can make mistakes, so important compliance decisions should be checked.

Why is it a beta?

We released Nova early on purpose. The questions teams ask in their daily work show us where the assistant helps most, and the feedback from the beta shapes what it learns to do next.

Why it matters. Answers that used to need a report or a colleague are a question away.

Agentic AI

Agents that work cases on your rules

The problem. Last month AI Suggest started proposing decisions, but someone still had to apply each suggestion, even for cases where the team's policy was already clear.

What changed. Agentic AI acts on cases in the background, following rules that you design, starting with name screening.

How do you build an agent?

Open Agentic AI under Monitoring Rules in the left menu. An agent is a flow drawn on a drag-and-drop canvas. It starts with the AI's suggestion for a case, then passes through conditions you set, then ends in actions. Conditions can use the AI's decision, for example true positive or false positive, and its confidence score. A condition such as "if the AI decision is False Positive and confidence is above 90%" leads to one action, and the remaining cases to another.

What can an agent do?

Actions include closing the case, changing the risk level, assigning the case to a specific user, adding the AI's reasoning to the case as a note, and calling your systems through a webhook. Branches let one agent treat clear cases and uncertain ones differently. Clear namesakes can be closed with a note that explains why, while uncertain cases go to a senior analyst with the AI's reasoning attached.

When does an agent run?

An agent can run on demand or on a schedule, for example daily or weekly, so it keeps working through the queue in the background like a digital colleague. The team decides which cases an agent may touch and what it is allowed to do with them, and every action it takes is recorded on the case.

Why it matters. Decisions the team has already made can be applied consistently and around the clock, and analysts spend their time on the cases that need judgement.

One platform for AML, KYC and fraud.

Screening, transaction monitoring and client risk on one engine, with one risk score for every client.

800+companies use complead
80+countries served

Transaction monitoring

AI Suggest for transaction alerts

The problem. AI Suggest helped name screening reviews, but transaction alerts, often the larger queue, didn't have the same support.

What changed. AI Suggest now works on transaction monitoring alerts.

How does it help with a transaction alert?

When you review a suspicious transaction, AI Suggest proposes a decision with a confidence score and the reasons behind it, drawn from how similar alerts were decided before. The analyst reads the reasoning next to the transaction details, then accepts the suggestion or decides differently. As in name screening, the decision stays with the person, and it is recorded as theirs.

Why it matters. Transaction alerts move faster, and the reasoning behind each decision is easier to follow.

Onboarding

KYC form templates and screening on submission

The problem. Building a KYC form from an empty page took time, and screening an applicant still waited for someone to start it after the form arrived.

What changed. KYC forms start from templates and can screen applicants automatically.

How do the templates help?

Open KYC Forms under Customer Management in the left menu and start from a ready-made individual or corporate template instead of an empty form. Adjust the fields to your process, set which ones are required and add validation rules so incomplete or badly formatted answers are caught before submission. Then share the form securely with your customer.

How does screening on submission work?

In the form's settings, switch on sanction screening on submit and choose the search profile and minimum match rate to use. When the customer submits the form, screening starts in the background, and the results are ready when the reviewer opens the submission. Uploaded documents can be previewed in place, and the submission keeps details such as the IP address it came from, which helps when the onboarding is reviewed later.

Why it matters. The first screening happens at the first contact, and the reviewer sees the answers, the documents and the results together.

Dashboards

Widget Builder

The problem. Ready-made charts answer common questions, but each team has its own questions about its own work.

What changed. The Widget Builder lets you design widgets from your own data.

How do you build a widget?

The builder works in three steps. Choose a data source, such as customers, transactions or cases. Add filter rules to narrow it down. Pick a visualization, such as a bar, line or pie chart. A Live Preview on the right updates as you change the rules, without pressing a button, so you see the result while you build it. When it looks right, add the widget to your dashboard in one click.

Why it matters. Each team tracks the numbers it cares about, in real time, without waiting for a custom report.

Everyday work

Smaller changes you'll notice during the day

  • Rule Simulator: a Simulator page under Monitoring and Rules tests rules on historical data, and the canvas can run a rule on a random transaction to show the result and how long it took.
  • Swift in transaction monitoring: Swift transactions can be included in transaction monitoring and scored by your rules.
  • Adverse media in bulk: names can be uploaded for adverse media screening through Excel.
  • AI summary on customer cases: the AI case summary is available on customer cases.
  • Why a risk score changed: the customer detail page shows which sources and rules feed the current risk score, and clicking a point on the risk chart explains the change.
  • Bulk open and close: customer cases can be opened or closed in bulk.
  • Live Transactions: the level filter allows several levels at once, and the list loads as you scroll.
  • Canvas tooltips: every node on the Canvas Rule Builder explains what it does.
  • Exports: batch exports include a Max Match Rate column and case exports include the Last Closed Case Reason.

If you already use complead, the details of each update are in the Help Center inside the app. For other months, see the Product Updates category.

Previous month: May 2026 updates. Next month: July 2026 updates.

See it on your own cases.

We'll show you Nova, Agentic AI and the Widget Builder with your own examples.

7-day free trialNo credit card required
800+companies use complead
80+countries served

Frequently asked questions

Can Agentic AI act without limits?

No. It acts only through the conditions and actions you draw on its canvas, and every action is recorded on the case.

Which module did Agentic AI start with?

Name screening. Other modules followed in later releases.

See the July 2026 updates
Does Nova see data from other organizations?

No. Nova answers from your organization's data, within the permissions of the user who asks.

Is screening on submission mandatory for KYC forms?

No. It is a setting on each form. If it is off, screening can be run from the submission detail.

Enes Keltepe
Written by Enes Keltepe Technical Product Manager Enes Keltepe is a Technical Product Manager at complead with a background in software development. He focuses on product development, software engineering and financial technology.

Originally published

Back to blog