Voice – It’s Still Good to Talk (And Someone or Something Might Be Listening)…

Published On January 28, 2019

A Behavox voice analytics special.

The history of voice in the financial markets is long and rich. It has been integral from the very beginning, central to the open outcry on many physical markets, the basis for the exchange of quotes and advice to ensure customers are understood, treated fairly and relationships are transparent.

When markets evolved from physical exchanges to voice-based dealing and then increasingly to the electronic ones that prevail today, the telephone was the instrument that connected parties. That evolution continues today, from mobile devices to more advanced forms of technology and social platforms. As the channels of communication proliferate, the potential for further development is boundless.

In a series of articles in this issue, Radar explores the status quo, the regulatory challenges, the technology now, and what the future might look like.

Former banker turned fintech leader Germán Soto, an investor and technologist, analyzes the financial market drivers that are shaping how software will change to meet very specific demands. Behavox R&D specialist Arseniy Gorin explains how academic and commercial research is benefiting the latest tech, and the challenges behind languages and the typically noisy multi-speaker dealing room environment.

The opportunity in audio is not limited to compliance (and extra cost) but is firmly focused on generating alpha (profit) and we cover the pioneers in this area too.

Is Voice the final frontier? It is certainly one that the final services market is now exploring.

The regulatory scandals that have dogged the industry post-financial crisis have had an extreme impact on behavior, as well as many of the new regulatory requirements. The Libor rigging and FX manipulation cases revealed that, in most cases, the archive and retrieval systems for voice were almost entirely inadequate at most of the banks and brokers that were subject to regulatory requests for information.

In some cases, there were no voice records at all; in others the banks claimed there were none, and then someone stumbled on them! Data quality was almost universally poor. Because of archaic systems, the cost of storage and the sheer size of the voice files, compression was the only solution, and this reduced quality in every instance. Those that followed the enforcement proceedings closely could see that voice records, where they existed, were far from the solid evidence required to secure any prosecution.

Attention turned swiftly to chats and emails; these soon became the ‘smoking gun’ that the prosecutors depended upon for evidence. But the wave of new requirements around audio recording and monitoring show that the regulators have not forgotten the importance of voice. The technology and costs for storage have improved massively since these antiquated systems and approaches were installed at the banks. Analytics have too.

The days of disconnected individuals, sitting in call centers in far-flung parts of the world or in separate states to the entity, listening to random phone calls with zero knowledge of the subjects, people, asset classes and regulations are over.

‘Done.’ – who on earth was that on the line?] graphic – mock up text as speech bubble and “who on earth…” as a separate thought bubble.