BEHIND THE INSIDER
Meet Kristina Haag, Head of User Success
Neterium is an innovative API-native company reinventing Financial Crime Prevention and responding to the explosion of market demand for highly performant, scalable and efficient watchlist screening.
Neterium’s ultimate goal is to allow clients and partners to achieve excellence in compliance while delivering frictionless customer service.
Built and developed by senior industry experts and top-notch technology specialists, Neterium’s team is unique.
80% of international banking transactions worldwide are screened today by solutions that the Neterium founding team had previously developed.
To spotlight the best insights of Neterium’s excellence, we ask a few questions to the inspiring crew behind it.
As part of a new series, Behind The Insiders, today we’re speaking to Kristina Haag, Head of User Success at Neterium.
1. Hello Kristina, can you tell us a little about yourself and Neterium's Head of User Success role?
Yes, sure. I have spent a significant part of my career in the screening business, both on the technical side, from Front-End Developer to Software Engineering Manager, and on the practical side as Business Analyst and Product Manager. I'm passionate and enthusiastic about the combination of data, linguistics and technology, so watchlist screening is the ideal way to do what I love every day.
As Head of User Success, my role is to help technical users maximize the value of our screening API. This means assisting and guiding product and engineering teams throughout their entire user journey, from their first steps in our sandbox environment, through their evaluation and Proof of Concept initiatives, integrating our API in their system until reaching a stable basis for day-to-day operations management, monitoring and support. Our ultimate goal is to enable and empower our users to master our API efficiently and continuously gain maturity and autonomy.
“My ultimate goal is to enable and empower our users to master our API efficiently and continuously gain maturity and autonomy.”
2. In a very competitive environment, the role of a customer's user success is not just about "winning the customer" but "showing the customer the path to value." What are your metrics and incentives to do that?
In the context of APIs, to win a customer, you need to get the adherence of the technical teams first, and they need to be convinced about the value of the product. At Neterium, we believe that this conviction can only be achieved by concrete hands-on testing and live experience.
As part of our onboarding, we have a particular focus on users and dedicate our exclusive attention and support to make sure that all their concerns are resolved before even the contract is signed.
On our Dev Portal, users can find the complete documentation of our API and access our sandbox to get a feel for the abilities of the screening engine. Training and Q&A sessions are organized to help users get the best out of our screening engine, address concerns and pain points, and align their needs with the screening engine's capabilities. Users can freely access our API for several weeks to evaluate all our features and perform a full Proof of Concept from all angles (effectiveness, efficiency, performance). You can access Neterium’s Dev Portal here: https://portal.neterium.cloud/
To measure the performance and quality of the model in the light of our users’ risk policies, we also have developed a unique Model Validation tool. It relies on "predictive accuracy" to measure how close the model's predictions will be to what happens.
A simple user interface allows uploading a data file, then provides a visual overview and comparison of multiple runs with different configurations and allows review of all test cases and scenarios in detail, highlighting conflicting or inconsistent expectations to suggest a new configuration for another run.
3. Model Validation and Model Risk Management are not recent trends, and in general, independence is a master word in this context. What is different in your approach, and how does it fit into this picture?
That's true. Model Validation has been a well-known requirement in the screening domain for quite a while. However, surprisingly, the aspect of independence is generally taken at face value and translates into costly consulting and third-party audits or model-independent test data generators.
At Neterium, we fully agree with the "degree of independence from model development and use" that the FED has underlined in its Supervisory Guidance on Model Risk Management (SR 11-7) in 2011. But we couldn't agree more with the associated concession that "independence is not an end in itself" and that "validation work might most effectively be done by model developers and users", as long as it is "subject to critical review by an independent party".
This is where our users and partners play a crucial role. Their Compliance departments are reviewing the development, design and use of the system daily, challenging decisions, providing valuable feedback on pain points and pushing for a continuous improvement of our system. Model Validation, consciously or unconsciously, is present throughout the entire life cycle of the screening engine.
“Compliance departments (of our users) are reviewing the development, design and use of the system daily, challenging decisions, providing valuable feedback on pain points and pushing for a continuous improvement of our system.”
To ease this primary and omnipresent task of decision-making at all stages of the screening project and efficiently close the feedback loop with our customers, we have designed and built this Model Validation tool that allows and/or assists users in defining their model (and expectations), review and validate this model with their Compliance Team, then use it in different phases and for other purposes during their project:
- Proof of Concept to assess how well the engine performs in terms of effectiveness and efficiency regarding the expectations.
- Tuning to perform and compare multiple runs to find the optimal configuration regarding the expectations.
- Acceptance Testing to assert that all expectations are met and all tests passed.
- No-Regression Testing to compare two runs and assert that screening results between two runs are identical or better.
- Periodic Reviews & Audits.
4. Neterium's innovative SaaS solutions combine the benefits of the latest technologies (e.g., Parallel Computing, Machine Learning) with in-depth financial crime compliance knowledge. False Positive Hits are a well-known pain point in screening engines, and Model Validation generally focuses on effectiveness. How can users make sure that your screening engine is practical and efficient?
Since the beginning, our screening engine has been built with efficiency in mind, and the out-of-thebox configuration provides a commonly accepted level of efficiency.
It goes without saying that real hits must not be missed under any circumstances. However, an effective screening engine that raises all expected alerts but floods them with false-positive hits leads to a very inefficient resource and time-consuming review process for daily operations. Moreover, with so many alerts to review, there is a high operational risk of ultimately overseeing a true hit.
This is why it is important to validate the effectiveness and include efficiency as a core part of the model. Our model validation approach brings relief as it integrates efficiency assessment and strongly links the two goals.
5. You shared with me that in your previous life, you were a postwoman. How do we go from delivering mail to Head of Success of one of the fast-growing Fintech companies?
It was a student job, but it connects with my today's role in some aspects. I often say I started as a Postwoman and ended up using Postman.
As a postwoman, you are expected to sort parcels and letters so you can prepare the delivery bags for your round. On foot, with the help of a trolley, you deliver the mail to homes and businesses, re-direct wrong mail, collect signatures…
And Postman is an API platform used for developer onboarding, exploratory testing, automation and visualisation. It's an essential part of my daily job.
“I often say I started as a Postwoman and ended up using Postman.”