SBS Blog > Financial communications idea will read between the lines by James Bowden

Financial communications idea will read between the lines

By James Bowden - Posted on 21 July 2026

Two academics from the Department of Accounting and Finance, James Bowden and Mark Cummins, have created a business idea they hope will help investors make the right decisions – and might even help identify fraud. Here, James outlines what ReadBetween is all about.

Our new financial communications venture ReadBetween aims to ‘read between the lines’ of financial reports to get to the true story.

The idea is to go beyond textual analysis and investigate the audio component of annual reports or earnings calls – conversations that firm managers have with investors – to pick out red flags in the tone or pitch that indicate there may be more behind what is literally being said.

My background is in text analytics in finance, using machine learning to try to identify patterns and trends in terms of what companies say in their annual reports, and what they say in earnings calls, the transcripts of those calls that they have with investors. I’ve always been interested in the audio component because when you’re just looking at text, you’re losing all that verbal information, in terms of how people have said certain things on the call, that provide insight into a person’s emotions or opinions.

Thanks to a Research Excellence Award (REA), we had a PhD student look at multimodal sentiment – that's text-audio emotions in financial texts. Since then, we’ve been doing a lot of work on the audio side – trying to measure cues such as hesitation and identifying deception. For example, research tells us that when someone speaks in a lower pitch, that’s generally associated with, or perceived as, being more confident, so we’re taking these principles and applying them to the finance and earnings conference calls. We’re looking at the hidden verbal clues that might tell us the financial picture is not as rosy as we’re being told.

Phonetic tools that pick up on speech rates, as well as variation of pitch and tone, can be used to identify potential issues and AI is also being used to interpret the output, looking for verbal red flags and abnormalities in speech which differ from the impressions the transcript alone provides. Those verbal clues can then be used by the investor to investigate the situation further. There is also potential at a later stage to use this new tool for fraud detection in a financial services setting.

For markets to run efficiently you need capital and money to go to the right places - to the companies that are genuinely growing, that are providing benefit to the economy, to society - and that requires transparency from corporations about their true performance and projections. What we’re trying to do here is help identify embellishment of a situation, or even fraud on an extreme scale, and ensure that funds are going to the companies that do genuinely provide growth and benefit to the economy and society.

As well as company performance, we can hopefully identify when companies are not being completely truthful in terms of sustainability - greenwashing for example - where a firm talks about its sustainability and its environmental performance. This potentially gives us a tool to try and identify that kind of misrepresentation as well.

We’re so grateful that we were awarded £25,000 thanks to the Stephen Young Entrepreneurship Awards at a recent Strathclyde Inspire event, which will help progress our financial venture. We plan to use this funding to develop the tool and back test it; this involves using historical information - where outcomes following conferences are known - to identify the vocal identifiers and test the accuracy of the model. The funding will also be used to judge the extent to which there’s a market for this information.

We know however that the costs of carrying out the historic research could be substantial; with thousands of conference calls and data to be analysed, that racks up a lot of costs in terms of time and access to specialist equipment such as a Graphics Processing Unit and ARCHIE-WeSt, a regional supercomputer centre.

Tools such as a GPU and ARCHIE will allow us to carry out the analysis much quicker and on a much bigger scale. We wouldn't be able to do that without the funds. It will allow us to bring in a developer who can push things forward a lot quicker. It’s quite exciting introducing an entrepreneurial aspect to our research, and hopefully, ultimately, we’ll be able to spin out Read Between as a result.