Since the summer, we’ve been shipping a lot of things our customers and Open-Source community have been asking for with one simple goal: help Compliance, Risk, and Ops teams adapt faster, investigate smarter, and help your tech & security teams keep full control over how Marble is deployed and accessed.
This fall product roundup brings you:
- New AI assistance for rule building and AI enrichment (OSINT)
- Adding AI customization settings to make our AI agents truly relevant to your organisation
- Workflow automation between Marble and your internal systems
- New capabilities for the Case Manager to help teams handling large alert volumes
- Enterprise-grade SSO and network controls for security-conscious deployments
1. AI to build transaction monitoring and AML Screening programs better and faster
AI Rule Translator – understand and validate rules in plain English
Evolving a detection program is hard: you’re constantly adjusting rules to new fraud patterns, products, and regulatory expectations.
To help with that, we’ve shipped the AI Rule Translator.
It takes any rule in Marble and explains, in plain English:
- What the rule is trying to detect
- How it fits into your broader detection framework
This makes a real difference for day-to-day work:
- Faster rule adaptation: quickly see what a rule actually does and its potential impact for a more rigorous creation and confident validation.
- Enhanced operational redundancy: facilitate seamless knowledge transfer and team collaboration on sophisticated rules, reducing reliance on individual expertise.
- Easier audits and internal reviews: share clear explanations of detection logic with regulators, auditors, and internal stakeholders, fostering transparency and trust.
This is the first step toward our longer-term vision: an AI assistant that not only explains rules, but also helps you continuously optimise and evolve your detection framework.
Motiva, an API to boost Screenings performances in high-concurrency environments
When a transaction is being processed, every millisecond of latency matters.
As our customer base grew and our processing volumes increased, we encountered performance challenges that led us to build a Customer & Payments Screening API specifically designed for high-concurrency environments: Motiva.
Deploying Motiva in production has delivered measurable improvements to our screening infrastructure:
Latency stability
- P99 latency went from 249 ms to 160 ms
- Maximum latency went from 1.26 seconds to 254 ms
P99 Latencies remain consistent even during traffic spikes that previously caused elevated tail latencies.
This reliability is critical for payment processing where delays cascade through dependent systems.
Memory efficiency
- Memory footprint went from ~50% of 4 GB to 3 % of 512 MB
- Memory consumption stability remains stable regardless of concurrent request count.
We no longer need to scale instances proactively based on anticipated load spikes, simplifying operations and reducing costs.
More efficient resource utilization means higher throughput on fewer instances. The operational cost savings compound over time as transactions volume grows.
🚨 The fight against financial crime requires both breadth and depth.
→ Comprehensive data from organizations like Open Sanctions
→ Performant infrastructure that can apply that data at scale
Motiva is our contribution to the latter, built with gratitude for those who enable the former.
👉 Get the full details on Motiva in this article: https://www.checkmarble.com/blog/high-concurrency-entity-screening
2. Improved usability and more AI Assistance for your investigations
This summer, we launched the AI Virtual Analyst: an assistant that does the groundwork on each case before your team even opens it — summarizing key facts, highlighting red flags, and suggesting next steps.
This fall, we’ve taken it further on two fronts:
OSINT enrichment
The AI Agent can now enrich cases with external data, helping you build a fuller picture without leaving Marble:
- Negative news and adverse media
- Potential PEP links
- Public job and company information
All of this is integrated directly into the case analysis, so investigators don’t have to jump between tools and browser tabs.
Customizable AI, aligned with your procedures
You can now tailor how the AI reviews cases to match your own risk appetite and workflows:
- Instructions for what to enrich (which sources, which signals to prioritize)
- Instructions for how to review a case (what to look for, what to weigh more heavily)
- Preferred output format and language for the AI-generated review
This means the AI Virtual Analyst doesn’t just “summarize a case” – it behaves like a junior analyst trained on your way of working.
A redesigned Case Manager for teams at scale
We’ve fully revamped the Case Manager interface to make it more comfortable adn easy to navigate. We also added features for investigators and managers to better work through large volumes of alerts.
Two areas of focus came directly from customer feedback:
- Advanced search & filtering
- Quickly drill down to the exact cases you need (by risk type, status, assignee, tags, etc.).
- Bulk actions
- Save time when updating multiple cases at once — especially useful for periodic reviews, reassignments, or mass status updates.
The result: better visibility for managers, and fewer clicks for investigators.
3. Workflow Automations, now talking to your internal systems
Marble already lets you automate work inside the platform:
auto-assigning cases, routing them by risk type or product, and triaging queues.
We’re now extending that automation beyond Marble, so your workflows can span across your existing systems.
Case Data Retrieval endpoint (Live)
You can now programmatically retrieve case data from Marble to trigger actions in your systems.
For example:
When an investigator tags a case as “Documents needed”, your system can automatically send a document request to the customer — without any manual copy-paste.
Case Update endpoint (Coming Soon)
Next, we’re completing the loop: your internal systems will be able to update cases inside Marble.
For example:
Once documents are received and verified in your KYC system, Marble can automatically update the case status, upload the document, ping a specific user, or escalate to another team.
Together, these two endpoints unlock true end-to-end process automation: Marble becomes the hub for risk and compliance decisions, while your existing systems execute and report back in real time.
4. Improved user management & security
For many of you — especially those in regulated or on-premise deployments — identity, access, and network controls are just as important as detection features.
This fall, we’ve doubled down on that.
OpenID Connect SSO – bring your own identity provider
Historically, Marble relied on Firebase for user authentication. It was great to bootstrap, but not ideal for on-premise and security-sensitive clients: it’s cloud-hosted, and it is yet again another layer to manage the authentication on top of their existing stack.
So we implemented the industry standard for managed SSO: OpenID Connect (OIDC). Now you can use dozens of supported providers, including auth0, Google Workspace, Microsoft Entra ID, Okta, WorkOS, Keycloak, and many more.
→ check the documentation for more details https://docs.checkmarble.com/docs/sso-openid-connect#/
And now you can manage all your Marble users in your current solution.
- Automatic MFA & Policy enforcement: your corporate security rules, including Multi-Factor Authentication (MFA), are automatically extended to our platform. We leverage the security measures you already trust.
- Single Sign-on (SSO): No more remembering or resetting passwords for our application.
One less dependency. One more security layer.
IP whitelisting – control where Marble can be accessed from
To help you further lock down access to Marble, we’ve introduced IP whitelisting for both:
- The Marble API
- The web interface
You can now restrict access to a set of trusted IP ranges or specific IPs — whether that’s your office network, VPN egress, or specific data-center ranges.
What’s next?
Everything above is available to help you:
- evolve your detection frameworks with AI support,
- make investigations faster and more consistent,
- give your security & engineering teams the controls they expect from a critical platform.
If you’d like a walkthrough of any of these features, or want to see how they fit into your existing setup, book a meeting from the button on the right !
