On private debt’s tech frontline

Could you tell us a little about the history of Accelex to date?

Founded in 2018 in London, Accelex provides data acquisition and analytics solutions for private market fund investors and asset servicers. The technology enables firms to have timely access to the performance and transaction data reported by those funds, in a clean and ready-to-analyse format. Powered by artificial intelligence tailored specifically for private markets, Accelex automates the extraction and preparation of data, allowing real-time insights from difficult-to-access unstructured data.

Your technology offers details of underlying investments. Can you explain how it works, and why it’s important?

Accelex’s technology is able to model and extract complex data sets from unstructured content – primarily PDF reports in the private markets industry – including financial and investment data pertaining to individual asset investments of private markets funds. To automatically read these documents and obtain that information we have to:

1. Render the document and identify textual and positional information inside of it.
2. Identify the various different elements that it is made up of (pages, sections, tables, charts).
3. Label those various elements and blocks of useful data using a standard industry-specific taxonomy (KPIs, funds, assets, currencies).
4. Link those elements together into a coherent network that fits the mental model of a private markets professional (eg what fund is linked to what portfolio company).
5. Finally, assess the results of steps one to four and determine what we should present back to the end user for validation and exception handling. We do all this with a proprietary engine made up of some of the best machine learning models available for various stages of that task, trained on a large variety of existing private markets documents. This is the brain of our platform, and the intelligence inherent in it is shared between all users of our platform. This gives our clients quick access to thousands of data points on the underlying assets and securities level, where they previously could only access just a fraction of the data due to the sheer volume of operational effort that would be required to obtain it by hand.

What do you see as the weaknesses in the way that private markets funds have traditionally reported to their investors?

There is a lack of reporting standardisation in private markets. Every fund manager reports differently, and the quality of reporting varies drastically across the market. In fact, it is difficult to talk about this space as a unified market from that perspective, as the kind of information an investor may expect from an early-stage VC fund manager is completely different to what they would receive from a blue-chip buyout fund. Consequently, building a unified view of your portfolio with the same level of insight at all levels could be virtually impossible.

This first issue is of course exacerbated by the prevalence of unstructured reporting to a fund’s limited partners, instead of providing packages of structured data. Some funds submit reports that are over 200 pages long to investors, with dozens of data-heavy tables breaking down multiple aspects of performance, cashflow positions and accounting. It is remarkably challenging for investors to use any of this information, especially when their exposure to private markets starts to grow beyond a few dozen funds.

How far has the application of AI/technology to private debt processes in general advanced?

We like to think that we have advanced it quite a bit over the last few months with the release of our private debt module for our clients. The additional complexity of private debt investments means reporting requires more granular data for investment monitoring and decisions. Debt teams need access to instruments rather than just asset-level information. In response, we have upgraded our data science capabilities to capture that detail.

To do that we have built an algorithm that identifies securities using all their underlying details such as maturity dates, coupon rates, credit spread, interest rate floor and the instruments’ seniority in the capital structure. This technology has much broader applicability than purely debt. For example, venture capital funds also report underlying investment performance on security level, often by the funding round, and may invest in common stock or preferred equity. This means that those clients who want that next level of analytics on a deal-by-deal basis are also able to benefit from our instrument-level extraction solution.

Where is human input still important in private markets data collection and presentation?

The idea of our product is to have a ‘human in the loop’. Investment professionals will always be required to make the right data-driven decisions for their businesses and customers. Our job at Accelex is to equip these people with scalable technology that takes away a significant portion of the burden of acquiring and normalising that data to generate the insights they need quickly and efficiently. Checks and controls around data quality will always be required to some extent; people need to have trust in the data they are relying on. Our software offers the benefit of automation but retains the confidence of seeing the data and where it came from. That confidence can be supported at the point of data acquisition using a variety of exception-based workflows.

How has the interest rate environment impacted private debt?

Private debt has exploded in size over the past decade. But the market has grown up in a period of exceptionally low interest rates and is now faced with the most extreme tightening in monetary policy since the late 1970s.

US interest rates have increased 4.5 points in 12 months; the fastest US tightening cycle in more than 40 years and twice as fast as the 2004 to 2006 hiking cycle that preceded the great financial crisis. The ECB’s rate hike is most extreme since it was formed in 1997. So far, the impact is muted, but they have been rising. According to Goldman Sachs the 18 defaults in US private loans year to date is more than for 2021 and 2022 combined. And the rule of thumb is that a rise in rates typically feeds through to credit with a 18-to-24-month lag.

It is likely that the rise is an extraordinary opportunity for distressed lending and for secondary purchases of private debt funds. Both as existing borrowers struggle to service their debt and as previously equity-only companies look for alternative funding sources where investors have reassurance around recoveries and their claims on the assets. And yet, investors into private market funds often have remarkably little information with which to analyse the sustainability of the debt burdens of their portfolio companies.

Nicole Weder is co-founder and chief product officer at Accelex, a London-based alternative investment-focused software company