In part one of this transcription-focused blog series, we provided a detailed overview of the industry’s progress in its quest for automated transcription perfection. In other words, a future where every word uttered between a trader and their client is captured at near 100% accuracy – jargon and all.

‍

Given as many as half of the respondents to our survey – which included managing directors from the likes of Morgan Stanley, JP Morgan Asset Management and Citi – felt concerned around the accuracy and reliability of transcription services, the snapshot was likely eye-opening.

‍

Several specialist automated transcription technologies have already proven more accurate than humans when listening to capital markets conversations. Moreover, thanks to dramatic enhancements in underlying AI and ML technologies, the level of accuracy is set to surge even higher over the coming months as more asset-specific transcription models are developed.

‍

In this part, we will explore just how transformative a widespread adoption of these transcription models could be for financial institutions, as well as a few examples of where firms are already seeing considerable strategic benefits.

‍

‍

Transcription aids tenacity

‍

Let's briefly revisit our construction analogy from episode one. Essentially, highly accurate transcription software forms the foundation on which two overarching strategic advantages can be achieved: future-proofed regulatory compliance practices and more data-driven front office operations.

‍

In the world of construction, robust foundations help to ensure a building holds up against a wide range of external pressures, be it an earthquake, hurricane, or tidal wave. For firms operating in the financial markets, one of the most significant external forces is regulation. New regulatory requirements may not strike as suddenly as a natural disaster, but they can be just as damaging to an unprepared firm’s balance sheet and reputation.

‍

This has been especially true over recent years, with authorities cracking down hard on financial institutions across all aspects of their operations. We need only highlight the scale of the SEC’s multiyear instant messaging probe, which has seen the regulator issue in excess of $2.5bn in fines to Wall Street banks over record-keeping misconduct. And watchdogs’ tails across the pond are wagging just as frantically. There remains upheaval in the City over the FCA’s recent ‘name and shame’ plans, whereby the regulator will publicly name companies facing an enforcement investigation on a more regular basis and at an earlier stage.

‍

With regulatory pressure growing and cost pressures biting – among a host of wider macroeconomic headwinds – deploying capital markets trained transcription technology could prove essential to safeguarding firms from significant financial penalties and reputational damage over the coming years. This is particularly the case given communications norms in the sector continue to evolve at an unprecedented rate.

‍

Those relying on subpar transcription capabilities are unlikely to remain resilient in today's tumultuous market environment, as economic headwinds rigorously test the integrity of their foundations. However, those that do have solid transcription foundations in place will find themselves in a very different position indeed. After all, only then can a building be fit with all the ‘mod cons’ that set it apart from the others on the street. In our case, only once highly accurate transcription tech is deployed can the data be used to enrich front office strategy.

‍

‍

Fuelling the front office

‍

Revisiting our client survey, it is interesting to note that when asked how relevant transcription services are to their role, over a third (37.03%) of respondents said they were unsure, felt it was not very relevant, or not relevant at all. Given the majority of respondents work in front office roles (34.78%), this may suggest a lack of understanding on trading desks around how transcription and natural language processing can enhance decision making.

‍

This is somewhat unsurprising, as there is a much wider array of advantages on offer than many realise. Once a firm can capture complex trader audio data accurately and – critically – ensure it is compatible and customizable with the firm’s broader internal data sets, a host of opportunities emerge.

‍

For instance, by processing a host of previous conversations had between traders and counterparties, a firm can create sophisticated buy-side insights reports. These enable the firm to keep track of incoming trade ideas and understand which liquidity providers are offering the best service, ultimately enabling the firm to identify the best counterparty for a particular trade.

‍

It also provides firms a clearer insight into which investment products are trending. By extracting product mentions, inquiries and quotes from conversations had across the business in real time and tallying them up, a firm can drill into what the market is asking for and what they are able to deliver. This could improve decision making for front office traders dramatically. Beyond that, companies can benefit from faster trade reconstruction and enhanced identification of new trade opportunities. And new use cases for natural language processing continue to emerge all the time.

‍

If all this talk of automation and natural language processing feels a little future-gazey, it is worth noting that several firms are already implementing these data-driven front office strategies. For example, one multinational Canadian bank is developing a large language model trained on text-based Bloomberg data. Once accurate transcription is implemented, this model could seamlessly apply to voice conversations, unlocking a wealth of insights for the bank’s front office.

‍

Indeed, the market leaders of tomorrow are already building future-proof foundations on which data-driven strategies can be forged. The importance of accurate transcription capabilities in achieving this cannot be understated.