Streamlining financial investigations with technology

Lakshmi Reddy, an Assistant Manager in our Forensic Accounting and Investigations team, talks about the benefits and capabilities of software tools used during financial investigations.

Whilst there is often no substitute to a detailed review of the numbers in a complex financial investigation, recent advances in technology can almost always assist at the planning stage of any assignment and can also assist as the investigation progresses.

In this article, we avoid the numbers, and instead focus on the various software tools that Quantuma regularly uses during a typical financial investigation.

IBM i2 Analyst’s Notebook (“i2”)

Prior to starting any investigation, it is important to fully understand the structure of a business and/or relationships between individuals and/or entities. In this respect i2 proves invaluable, both initially and during the investigation.

For example the chart below, showing relationships between entities and individuals, can be produced in seconds by importing the relevant company data from one of the many company information databases that are widely used or from information provided by the client.

As an investigation progresses, we can use i2 to quickly map further relationships, which can be anything from bank accounts to email addresses and phone numbers. For illustration, the following table was produced (again, in a matter of seconds) from a 30,000 line bank analysis to show the number (could just as easily be value) of transfers between various bank accounts included within the analysis.

To take this one step further, in the above example we can next seek to associate individuals/entities to the bank accounts by importing further bank data, perhaps obtained from the company payroll.

The requirements of any investigation will vary, but the key point is that i2 allows us to collate and visualise data, both structured or unstructured and of any size. It can quickly identify individuals/entities, connections, relationships, events, patterns and trends which, in turn, allows us to better plan investigations, both at the start and as they progress.

CaseWare IDEA Data Analytics (“IDEA”)

Another tool in our armoury is IDEA, which can be used to help plan and focus initial investigations.

IDEA is a commonly used audit tool that can ingest and analyse vast amounts of data, either from exported data (e.g. excel spreadsheets) or by connecting directly to a database (e.g. an accounting system).

IDEA has integrated, pre-built, standard audit tests that it can automatically run against the data in an accounting system to quickly identify potential anomalies for further investigation, and also allows bespoke queries/tests tailored to a specific case.

This enables us to quickly identify trends, patterns, outliers and duplicates in the data and also spot irregularities by analysing digits in numerical data (Benford’s Law tests) and to quickly identify larger value or round-sum transactions. In addition to spotting anomalies, IDEA also allows correlation, time and trend analysis to forecast and extrapolate data to help detect gaps and duplicates in sequenced data, for example, invoice numbers.

Altia Financial Investigation Toolkit (“Altia”)

When it comes to financial analysis, it is almost always preferable to analyse financial information within a spreadsheet, allowing greater accuracy, easy scheduling, reconciliations and comparisons.

Altia automates the process of converting paper-based information, by using advanced Optical Character Recognition (OCR) technology combined with complex back-end software, employing various logic checks on the converted data.

This allows us to quickly and accurately convert hard-copy records, such as bank statements or other lists/data into electronic spreadsheets for accurate analysis and, potentially, for importing into i2. It also allows us to ensure consistency in complex investigations where several team members are involved, as we can set up export templates to ensure that all data is exported and collated in a consistent format relevant to the matter at hand.

Altia is a valuable and time-saving resource that we use on the majority of our investigation cases, enabling tasks that would otherwise take several days, weeks and even months of manual keying to be completed in a timely manner. Electronic document conversion also means we can avoid the inevitable errors that arise with manual data entry.

Reveal Brainspace eDisclosure platform (“Reveal”)

Any robust financial investigation needs to go beyond the ‘traditional’ accounting systems. Inevitably, emails and other forms of electronic communication will contain financial information and/or communications that might be relevant to any investigation.

A full forensic investigation of computers and/or mobile phones is often needed and part and parcel of our offering. However, as an accountant, interpreting and understanding the vast amounts of data spewed out of forensic technology software can be confusing and often, in the first instance, unnecessary.

It is, however, important that an accountant has easy and focussed access to the financial data as part of the investigation, as they will be far closer to the financial data than a technology expert. For this reason we rely on Reveal, an eDisclosure platform which employs Augmented Intelligence (“AI”) to help focus investigations by enabling efficient review of data, emails, messages and files.

In addition to efficiently searching and reviewing documents using keywords, tagging anything that is relevant, and exporting relevant documents in a form that could easily be used and referred to in evidence, Reveal offers advanced investigative analytics.

Without any input from the user, Reveal employs AI to ‘learn’ from the user – Machine Learning. This means that as documents are reviewed and marked as Relevant or Not Relevant, the software analyses the contents of each reviewed document and then uses this information to assign a ‘score’ to all other documents to predict whether they might be Relevant or not. The reviewer can then target the documents with the highest scores with accuracy of predictions increasing naturally as more documents are reviewed.

Another automated feature of Reveal is Sentiment Analysis, which is independent of any reviewer actions and can help target initial investigations. The system analyses words and their correlation to one another, plus indicators of sentiment such as conspiration, intent, happiness, and anger (including overuse of upper case characters), to return email conversations of interest – identifying and categorising email/email threads as various emotions – Opportunity, Pressure, Rationalisation, Positivity and Negativity.

A user can then drill down into any category, drilling down by Domain, Topic and Users to easily find potentially relevant emails (illustrated below, using the publicly available Enron data):

The system will also automatically categorise emails by theme or topic, and represent this graphically – known as the Brainspace Cluster Wheel, as illustrated below:

This allows us to quickly identify and test relevant items, and to identify emails associated with these – drilling down into a particular cluster in the Brainspace Wheel shows an expanded view of that cluster and associated themes/topics within that section, which makes it very easy to click in and out of categories and therefore to identify potentially relevant emails for further review.

In addition to using AI to categorise and/or help identify emails that contain specific keywords or that can be categorised by reference to a certain subject or emotion, Reveal also uses AI to identify anomalies, i.e. emails or patterns that stand out as unusual when compared to the data as a whole. This can include out of character communications, code language or patterns and can be as specific as highlighting that someone sent an email to say “call me”, when that wouldn’t normally happen. It also identifies potentially suspicious gaps in communication frequency.

Summary

It is not possible to list all software that might assist with an investigation within the scope of this article, nor the complete capabilities of the software we have referred to. Financial investigations must now go beyond accounting systems and interviewing relevant individuals, with the use of technology a necessary component of any assignment. Technology will, of course, continue to advance and we can only imagine what we will be writing about in 12 months time!

Quantuma’s Forensic Accounting & Investigations team regularly undertakes detailed financial investigations, instructed directly by the client or by lawyers. The software tools referred to above are all available to us, we are trained in and regularly use them to assist our clients.

Lakshmi Reddy is a qualified accountant (FCCA) with significant experience in accounting systems and financial investigation software such as SAGE, Xero, Quickbooks, CaseWare IDEA and Reveal. She specialises in investigations, valuations and quantum.