Convier

USD 7,000 to 8,500 - That's the cost of each SAR being submitted to the Norwegian FIU. What’s the return on that cost? ?? Tune in to our latest podcast episode, where Carl Olsson from Danske Bank explores the effectiveness of AML regulations.

Imagine banks lending 1,000 skilled professionals to law enforcement for a year—working together on investigations and fraud prevention instead of each institution working on their own. A bold idea? Absolutely. Which is why we love it! ?? Listen in to our new episode with Carl Olsson, Head of Financial Crime Norway from Danske Bank. Link below ???

There will be more regulations - not less That’s the key takeaway at "Hvitvaskingskonferansen," the annual Anti-Money Laundering conference in Norway. But here’s the question: will adding more regulations actually make us better at fighting financial crime? Financial institutions are already drowning in false positives—mostly due to poorly implemented technology, but also because regulations require extensive reviews of each alert. I fear we will be needing more “False Positive Investigators”, but I hope I’m wrong! Want the same t-shirt? Comment “False Positive Investigator” below, and I’ll send you one! Convier

Hvitvaskingskonferansen er i gang, og i Convier tar vi hvitvask bokstavelig! ?? S?, hvis du trenger ? ?rense opp? i AML-strategien, kom innom og f? med deg en dose ren samvittighet (og litt vaskepulver)! Vi lover en prat som er like frisk som t?yvasken din ?? Kaja Kvello | Andreas P. Engstrand | Finanstilsynet (FSA Norway) | Finans Norge | ?kokrim

?? The worst and most time-consuming part of AML? Reporting. In AML, every piece of information needs to be sourced, documented, and ready to defend at a moment’s notice. And often, it feels like half the job is just keeping up with the paperwork. Reporting shouldn't have to be a burden. When your data is connected in a knowledge graph, adding new details or creating reports becomes effortless. No extra steps, no fragmented systems— it's all just there. Focus on the insight, not the paper fight.

How well do security professionals collaborate across sectors? Hedvig Moe |?Tor Indst?y | Andreas P. Engstrand

What happens when the lines between state and private threats blur? ?? As the dynamics between state actors and private entities evolve, the landscape of security and intelligence becomes increasingly complex. In our latest episode of Conviersations, we delve into these shifting boundaries and discuss how they impact our strategies for risk management and threat prevention. #Security #Intelligence #Geopolitics #Conviersations

What happens when criminals start using artificial intelligence? ?? How much data does a Chinese electric car actually leak, and how can we work effectively with threat intelligence when the lines between state and private actors blur? ??? The first episode of Conviersations is live! Andreas P. Engstrand interviews Hedvig Moe, Partner at Thommessen , and Tor Indst?y, Head of Threat Intelligence at Telenor. Tune in for insights on how geopolitics and technological developments impact corporate security and discover the strategies these experts believe are essential for meeting tomorrow’s threats ?? Available on Spotify and Apple. You know where to find the links In Norwegian

Something exciting is brewing at Convier... ????? Our office is prepped, the gear is ready. Stay tuned for more!

?? Curious about using LLMs in investigations? You might be surprised at how powerful they can be — especially for comparisons! If you give an AML investigator an empty AI prompt, their first question will likely be “So, who are laundering money?” It’s a compelling question, but not one LLMs are designed to answer. They're not trained to spot complex transaction patterns. However, where LLMs (even small ones that can run on your laptop) really shine is in comparing data. In its simplest form, you can ask, "Which number is bigger?" and get the right answer. But what if you could scale this to compare dozens, even hundreds of data points? That’s where things get interesting — and useful. ?? To make this work practically, you’ll need a strategy for data collection and preparation. The illustration below is how Convier does this, but the same principles should apply to other tools. There are a couple of key tricks to pull off: 1?? Selecting Relevant Transactions: Whether it’s from a pre-configured filter or inferred from the prompt, narrowing down the time range and transaction details is essential. 2?? Choosing the Right Groupings: Think about factors like country or month, and aggregations like sum, min, max, and averages across key fields such as date and amount. These can also be set up beforehand or interpreted from the prompt. Once the aggregated data is ready, convert it into clear, simple text that the LLM can process. The output? A detailed breakdown of the similarities and differences that could uncover valuable insights—ones that might have taken much longer to find manually