Ardian senior execs: standardisation in GP reporting is one of the largest barriers for AI

Could artificial intelligence change the game when it comes to LP-led secondaries transactions? Secondaries Investor caught up with Wilfred Small and Arnaud Mercier, managing directors at Ardian, about the technology it has been creating to help it monitor the performance of its portfolio and inform its underwriting.

What is Ardian’s approach to AI, data and technology?

Wilfred Small: We’ve been in the secondary business since 1999, so for over 20 years, and it’s always been in our DNA to have a data intensive approach to an industry that has information barriers, unlike the public market.

Today that database amounts to over 1,600 underlying fund investments with over 600 GPs and 10,000 plus underlying companies. The number of distinct data points that we’re getting from all of that information is over 4.5 million data points each quarter.

The insights that you can draw via artificial intelligence are only as powerful as the quality of the data and the breadth of the data that you’re able to digest and feed into the system. It takes time to build up the necessary relationships and history of data points that we have done. We think that’s something [that], despite technological disruption that we see today, remains a barrier to entry in the secondary market, in particular, on the large end for LP portfolios.

More recently, in particular, we’ve had two key milestones in terms of the development of our data team and the integration of artificial intelligence into our process that started in 2020. We developed a proprietary tool called iBIP. That’s a software tool that we developed that is able to automate some degree of the cashflow projections that we run on our portfolios. All of the underlying fund-by-fund and company-by-company modelling is done by our team. It’s informed by the interactions that we have with our GPs, but once we have that raw go forward projection as it relates to aggregating and sensitising and running scenarios across various different Ardian master funds, this tool helps aggregate that data and make that process more efficient. It’s also able to extract information as it relates to aggregate sector level valuation trends, by quarter leverage trends.

As a second step in 2022, we built out a dedicated data science team. It’s a dedicated team with five people in total, and we have people within our secondaries and primaries team within those five that are focused exclusively on this project. They work on iBIP in terms of developing the software and also future new programs and serving as the liaison between our investment teams and our dedicated data team as well.

You mentioned that Ardian has data from over 1,600 funds. When you’re looking at acquiring portfolios on the secondaries market, what percentage of those portfolios is there overlap with your existing database?

WS: Typically we’re an investor already in 80 percent or more of the funds that we’re acquiring, and so the whole philosophy is to buy what we know. We think it leads to differentiated levels of conviction and associated proactivity and speed with which we can approach sellers. From a risk standpoint as well, we have a very high degree of confidence as it relates to the underwriting and the base case that we’re making. That’s a huge part of our strategy: buying very high degrees of coverage.

Is there a standardised way that the GPs report their quarterly NAVs which helps the process?

Arnaud Mercier: That’s really one of the barriers today. Every GP has its own standard and level of information that they are able to share. There has been some progress over the past years with GPs trying to share more and being more transparent. Having said that, it’s very often through a PDF. It’s not like you can just take the information and put that into a database and generate your extraction. That’s always the real problem for us. How are you sure that the quality is there and the data is accurate? When you get an EBITDA multiple, [whether] you ask the seller or the buyer, it’s going to be two different metrics. So the question is always, okay, which data [do] we think is accurate and can be used for us in terms of investment decision making.

What other parts of the process do you see where technology could be used?

WS: As it relates to data aggregations, certainly a more systematic way of reporting would be good for the investor community, the community broadly and enable us to more effectively aggregate and extract data to develop our own insights.

The reality is that the private equity industry and the nature of relationships between LPs and GPs is highly nuanced and highly complex. There are strategic relationships and we don’t see that changing, particularly in the area of the market where we operate at the very large end. We’re oftentimes buying several hundred-million-dollar commitments in funds, even possibly larger, up to $500 million. When you’re focused on that size, again, the nature of the discussion both with the seller and with the GP as it relates to transfer is highly strategic in nature because it necessarily entails the transfer of a historical strategic LP relationship [with] the GP… we don’t see that changing.

The private equity industry still remains a people industry and the insights that we develop are only as good as the quality of the data that can be put in the system, which takes time to develop.

Wilfred Small is a senior managing director and heads up Ardian’s San Francisco office. Arnaud Mercier is a managing director in Ardian’s London office.