How artificial intelligence is supercharging asset recovery

Asset recovery, encompassing legal claims as well as recovery of traditional assets such as cash, investments and real estate, can often be likened to trying to find a needle in a haystack. It can certainly feel like that when you have spent weeks sifting through emails, WhatsApp/chats, banking records, contracts and all the other materials needed to unearth leads, trace funds and gather evidence to recover them.

As investigators and asset recovery practitioners, what we are up against when seeking to track down and recover assets is not just the weight of information, but also a deliberate effort to conceal and disperse the assets far and wide.

Faster and better

So, imagine the advantages of being able to deploy cutting edge tools that can help us carry out a lot of this painstaking work – not only faster, but also better.

For us here at Quantuma, the new weapons in our eDiscovery and investigations armoury include ‘Ask AI’ within the Reveal-Brainspace platform, having become one of the first firms in the asset recovery field to deploy this generative AI enabled tool.

The new ‘Ask AI’ functionality allows us to literally ask a question of the dataset to test out theories, or quickly determine the correct position and direction of the case. Being able to ask questions such as “What investments are linked to Company X between 2017 and 2020?” or “Which individuals were involved in the merger discussions – provide in timeline” or “list large money transfer discussions” saves a considerable amount of time in the initial information gathering phase and in getting to the important points. Experience also shows that this kind of insight naturally leads to further avenues of investigation, culminating in more detailed searches and a faster understanding of the issues at hand.

In turn, such next generation tools equip us with additional searching power where keywords alone may not be sufficient. Further benefits include an almost immediate summary answer, in an easy-to-read natural language, before carrying out the necessary, but time consuming first pass relevance review (which will naturally become more targeted due to the use of the tool).

Delivering in practice

Working closely with our legal clients, we have seen how AI can help to deliver outcomes that might otherwise not have been possible when using traditional techniques.

In a recent example, we were investigating the award of construction contracts to familial contacts. We expected the AI to support the identification of a corrupt procurement process and reveal how contracts were awarded, despite a government mandated “value-for-money” process. Whilst that is exactly what we found, the additional discovery of, what was in effect a Ponzi scheme, was unexpected when no related search terms were used, nor was a review undertaken into this issue, but this discovery was illuminated by the Reveal-Brainspace platform’s summarisation and visualisations capability.

In another case, we were trying to track down a crypto fraudster who had absconded with the funds from an exchange. An anonymous source sent us a photo of the fugitive in a public setting, without any context. We used open-source-intelligence (OSINT) tools to link the image to locations and key individuals, alongside combining previously used email addresses to social media accounts and online trading accounts. The results ultimately led us to the fraudster.

Additionally, there was a case where a large multinational was concerned about cash leakage but wasn’t sure where or how to look. We utilized the ‘sentiment analysis’ feature across a number of individuals and immediately found ‘negative emotion’ conversations between particular employees, pointing to culpability in fraudulent payments.

A further example was that whilst investigating discrepancies in interest rate swaps, we ‘trained’ the machine learning capability with examples of ‘good’ and ‘bad’ documentation, in order to be presented with documents that had a greater than 98% likely relevance ranking, leading to faster interrogation of the dataset and pinpointing the traders involved.

Game-changer for legal teams

The lawyers we work with are also seeing the benefits of these new tools. According to a Litigation Partner at a global law firm based in London: “The ability to test out our issues and get a feel for the client data, right at the start, is invaluable… it doesn’t replace review, it just gives us information faster.”

Further feedback was received from the Head of Litigation at an offshore law firm, based in the Cayman Islands: “We anticipate this being the standard in all our cases. It’s the perfect way to augment our investigation tool kit, and ensure we are looking in the right place”.

Strategically, these AI-enabled capabilities allow smaller legal practices to punch above their weight, taking on more data-heavy and complex cases without the need to hire more lawyers. For larger firms, this is an opportunity to do more for less, helping to drive case flow, while sustaining competitive fees.

Care and guidance

However, AI isn’t yet a silver bullet, nor has it completely replaced human intuition, insight and expertise.

Rather than being sentient, AI needs to be trained, guided and prompted by human input. We describe the current level of AI capability as ‘Augmented’ Intelligence rather than ‘Artificial’ Intelligence.

How the outputs are used – responsibly or not – and the benefits we can achieve is therefore down to us. The need for care and guidance in the use of AI is especially critical in the legal field, not just in ensuring that source data is reliable and setting the parameters and prompts, but also in harnessing the outputs to enhance the prospects of success.

Supplement rather than replace

Whilst AI in legal disputes and asset recovery has been around for many years, its adoption has been slow, due to misunderstandings around how far the technology can go and how much it might cost. In particular, some have doubted its effectiveness, while others believe it will replace traditional reviews.

What we are seeing now is a wider acceptance that AI-enabled tools are there to supplement, rather than replace traditional reviews, which has allowed newer technologies to be assessed and trialled. We are now at a junction where the technology is understandable (and not ‘black box’), and the users are keen to explore how this can speed up and improve their understanding of a matter, or lead to more successful avenues of investigation. This will of course improve the chances of success and save clients time and money, whilst improving asset recognition and recovery.

Watch this space

Looking ahead, crucial developments include improving connectivity between these AI-led platforms and OSINT platforms. In our view, the development of a ‘one way’ enquiry to OSINT platforms that doesn’t leave a trace, could enable immediate connections to be identified between, for example, individuals and their directorships and shareholdings, or real estate and motor vehicles, with their owners across multiple databases and jurisdictions. Currently, we have to employ multiple tools, so having better interconnectivity would result in significantly increased efficiency for asset recovery practitioners.

See for yourself

In the meantime, our best advice for lawyers is to learn more about the AI possibilities for asset recovery and try them for yourself. We are now providing short and targeted demonstrations of applications specifically geared to the issues and strategic plans of asset recovery for legal practices. If you would like to see how these AI-enabled tools will benefit your cases, please get in touch.