Sentieo Q&A: research workflows still using “outdated technology stacks”

AI thinking

Productivity provides an edge in equity research, and as equity-focused hedge funds navigate the markets this year, any opportunity to use next generation technology tools to trade more adroitly will likely be welcomed. In the following Q&A with Hedgeweek’s managing editor, James Williams, Nick Mazing, Head of Research at Sentieo Inc, discusses how its platform is using the latest AI technology to improve the research workflow and reduce the time taken to ingest structured and textual data sets, and empower portfolio managers as they look to build alpha-generating investment theses.

HW: Could you begin by providing some thoughts on what productivity can mean to equity analysts, when using new technology tools to improve research workflows, in light of the whole AI/machine-learning revolution?

NM: The adoption of new technologies is a competitive imperative not just for professional investors but for businesses in general. This has fuelled the SaaS and AI booms.

Today, we play an important role in bringing these applications to investment management. Our clients are seeing true progress: improving productivity, accuracy, and speed to insights.

Together, these improvements lead to better investment decisions. For example, our NLP and machine learning-based transcript summaries via email means trained models push information that was previously only accessible in linear format right to your inbox and enable you to read guidance language from the transcript right on your phone.

Another example of push is saved search alert, enabling you to get an alert every time the CFO of eBay mentions Amazon on call, or cast your net much wider with ML-generated synonyms in your searches. Having these tools, along with market and financial data, alternative data, and a purpose-built full research management system, can give you “superhuman” skills - and give your research an edge.

HW: What are some of the existing issues with research workflows that make it hard for equity analysts to make meaningful improvements (in terms of time, fewer manual tasks etc.), especially for those who are looking to avail of non-traditional data sets?

NM: The most common issue with research workflows today is an outdated “tech stack.” The question here is, when is the last time your firm evaluated the technology it uses, top to bottom?

Let’s start with a simple example: how many logins do you have to do every morning?

Technology has improved this issue a great deal over the last decade. The overall identity management has improved dramatically at the enterprise level, providing management teams with accurate insights on usage and engagement.

Now move this question to your workflow. Which parts of your process are done just because this is how they were done ten years ago?

If you have ten separate applications for your work, let’s say document search and reading, note taking, email flow, market data, financial/reported data, alternative data sources, and thesis management, you are wasting serious time for two reasons.

First, you’re wasting time switching between apps, and second, you’re missing the opportunity to augment your workflow through things like seamless collaboration and alerts. There is a lot more: for example, our interactive financial models let you enter your own assumptions so that you can quickly pre-qualify or dis-qualify ideas.

On the alternative data side, we combine data sets in the most predictive combination, and we have this across the system. The data can be screened across many metrics, it can be visualized and combined with other sets, and you can get push alerts for inflections. Alternative data is less about having “it” and a lot more about properly applying it.

HW: How can AI tools, such as the Sentieo platform, best be integrated into research workflows to aid equity analysts when it comes to uncovering fresh new insights that may not previously have been attainable?

NM: The challenge that all knowledge workers face today is too much information. This introduces two risks: missing key information, and not getting insights; all due to the sheer volume that one has to process.

This is where our ML and NLP tools come into place.

For example, we extract, classify, and rank transcript sentences based on classifications, such as guidance, and we overlay a sentiment and deflection filters on top. So rather than having to read a full transcript, you can see exactly what you need regarding guidance, and the NLP layer ranking sentiment and likelihood of “deflection”. You can also see terms and sentiment trends over time for better context, meaning you no longer need to listen to the five previous calls to get up to speed on “the tone.”

Another application is our very sophisticated synonym search. With synonym search, you don’t need to know whether a company uses “sales” or “revenue” to find because Sentieo understands that those are the same thing. This effectively cuts search time by 50 per cent.

Further, our ML tool will suggest autofill and similar search terms which ensures you see the most logical queries in a comprehensive way.

HW: Could you provide an example or illustration of how one of your clients, using Sentieo, has benefited from productivity gains in their research process?

NM: During a recent webinar “Productivity as an Edge in Equity Research” we had a customer from a large long-short fund discuss how they use our integrated Research Management System (RMS). Another customer on Twitter called us the “finance vertical version of Slack.”

The gains for both of these clients revolve around the benefits of having a purpose-built RMS that is integrated into your workflow as the central depository for the firm’s IP. Not only can you seamless read, highlight, tag, and share, but you can also keep your notes and screenshots and models, and any other files in one place, ready for collaboration and review, with proper version control, automation (such as having industry newsletters coming to your Notebook), and searchability: all available on mobile.

HW: Please comment on the importance of using systems/platforms that can support a wide range of research capabilities, and what you regard to be the biggest advantage of doing so.

NM: The investing world is always changing: this is why having a platform that is built for constant change is critical. There are data sets now that did not exist a few years ago: how have you integrated these in your work? There are substantial improvements in areas like screening across financial and non-financial metrics: how are your current systems accommodating these?

Dynamically improving search technology is also critical as new themes and topics emerge. For example: suppose you are looking for stocks with diverging transcript sentiment between management and analysts, with inflecting alternative data, and high FCF (Free Cash Flow) generation. Having these previously disconnected datasets combined with screening and other workflow capabilities expands everything an analyst can do in new ways.

HW: Finally, could you highlight some of the wider benefits of an effective research management program, underpinned with AI tools; such as, wider collaboration across global teams, securing IP, improving data governance and compliance?

NM: There are several organizational layers involved in research management systems. Today, analysts want a system that is comprehensive, easy to contribute to, and easy to search.

Portfolio managers want productive analysts and want to keep everyone on the same page, but they also want the ability to monitor the team’s activity from anywhere. More importantly, PMs need to ensure that the company IP stays with the firm as staff moves on. Having a proper depository with powerful search on top is a must. The firm’s technology professionals might be concerned around security, with topics like single-tenant hosting being front-of-mind.

Finally, compliance needs to have access to all work that has been done, with strict version control so that potential regulatory audits can be completed promptly. Using makeshift systems is still the industry norm. We see IP lost in shared drives, we see lack of basic searchability, we see consumer-grade apps in use by financial professionals. We are here to change this.



Author Profile
James Williams
Employee title