From Educated Guesses to Intelligent Appraisals: Generative AI in Alternative Asset Valuations?

The complexity in accurately valuing alternative assets presents a well-recognized challenge across the industry. Unlike publicly traded stocks or bonds, many “alts” are illiquid, and therefore do not carry readily available market prices. For example, a private equity investment in a large renewable energy infrastructure project may generate periodic cash flows, but placing an overall valuation on the asset can be challenging because the equity itself is rarely traded. Moreover, any particular project can have unique growth potential, operational challenges, or regulatory risks that are hard to quantify.

When attempting to mitigate alternative asset valuation problems, analysts often use strategies like mark-to-model valuations, discounted cash flow analysis, and net asset value (NAV) estimates. Another common exercise is scouring public market data for relevant comparable assets (comps). If the alternative investment shares characteristics with certain publicly traded assets, and you have the necessary time and analytical resources, then you can probably generate a reasonably good valuation.

For example, if your alternative investment is in a privately held commercial real estate portfolio, it might be possible to refer to a publicly traded REIT with similar property types and geographic exposure. Or, for an alternative debt investment, such as a private loan to a business, referring to public bonds or corporate credit instruments with similar risk profiles, industry exposure, and maturity structures would make sense. But the comparison still largely remains apples to oranges — an alternative asset compared to the standardized valuation regimes of traditional assets.

Moreover, these methods can be extremely time-consuming and resource intensive — and still leave asset managers dependent on opaque, infrequently updated valuation estimates. But what if you could just type a query into your own, custom-trained generative AI agent?

“Can you please provide an intra-quarter mark of my portfolio's renewable energy private equity valuation using the most recent private market comps available?”

Assuming your AI understood the composition of your portfolio and had been trained specifically on alternative investment valuation methods, you might be surprised at how good of an answer you got — and how quickly. By leveraging its capacity to aggregate and process exceedingly large datasets, including real-time data from private market transactions for assets similar to what you are holding, your AI could provide you with very accurate, and up to date, valuation estimates. Anytime you asked.

Generating intra-quarter marks with private market comps?

There is no doubt that AI is generating huge interest from alternative asset managers. A 2024 survey conducted by Alternatives Watch found that 78% of global asset owners believe that AI can help their teams make better alternative investment decisions, particularly by digitizing documentation and streamlining internal processes. Additionally, 42% of respondents already using AI said they have applied it specifically to portfolio analysis and modeling. These insights align with the growing sentiment across the industry that AI can be a powerful tool in valuation.

A report from AIMA, which surveyed hedge funds, noted that 86% of respondents permit their staff to use AI tools for operational and analytical tasks. Some firms are even developing custom AI models designed specifically for their portfolios. For example, Man Group, the world’s largest publicly traded hedge fund, has developed "ManGPT” — an in-house AI tool built by the firm's engineers tailored specifically to suit their portfolio management needs.

Aman Soni, VP of data strategy at Canoe Intelligence, on a recent Funds Europe podcast, asserted that collaboration between AI systems and human expertise will be crucial in adopting a data-driven approach to Alts valuation. Soni expressed that generative AI's ability to provide more frequent and accurate valuations could significantly enhance risk assessment in private markets. He even provocatively suggested that generative AI could be used to aggregate and analyze assets “with similar characteristics over time, potentially allowing for intra-quarter marks with private market comparables.”

Artificial intelligence: Hype or a real technological possibility?

Generating intra-quarter marks with private market comps would represent a significant shift from what is currently possible. While there is no concrete evidence of this happening yet, there are several potential pathways through which AI might be enabled to access the necessary information to provide intra-quarter alternative asset valuations, including integration with alternative investment data platforms.

There are companies, like Alai, that are already working on AI tools that can leverage proprietary data sources within asset management firms — and no doubt, many asset management firms are also investing heavily in AI internally. In fact, in spring of 2024, Mercer published a survey that indicated 91% of asset managers are currently (54%) or planning to (37%) use AI within their investment strategy or asset-class research.

A possible next step would be AI integration with platforms that track private transactions and valuations, such as Preqin (recently acquired by Blackrock), PitchBook, or Burgiss Group (recently acquired by MSCI). These platforms already aggregate data from private equity, venture capital, real estate, and private debt transactions. By potentially connecting to these via subscription-based APIs, an AI could tap into up-to-date valuation benchmarks for comparable assets.

AI could also be connected to data from digital platforms enabling fractional investments in alternative assets, such as Moonfare, a digital platform that provides education, networking, and unique access to “bite-sized” private equity investments for next-gen investors. Yieldstreet and Fundrise are two other next-gen platforms focusing on alternative assets and private markets. These companies likely have more frequent transaction data, because of their retail focus, and could theoretically anonymize their transaction histories and make them available through APIs for AI processing.

One other tactic for trying to generate intra-quarter marks with private market comps would be Natural Language Processing (NLP) techniques. NLP could be used by AI to scrape public documents, filings, and news sources that disclose valuations or transactions of private assets. AI models could be trained to extract insights from quarterly or annual reports, investor updates, regulatory filings, or even news articles covering private transactions to enhance valuation models.

The hurdles are real

Each of these methods could theoretically overcome the current data opacity in alternative assets, and make it possible for AI to generate accurate, near-real-time valuations. However, there are significant practical, regulatory, and industry-specific challenges that make it unlikely in the immediate future. Setting aside the standard challenges of preparing data in a consistent well-structured manner for any large computational project — and that alternative assets are inherently complex and heterogeneous — there are several other major challenges standing in the way.

One fundamental issue is the highly competitive, proprietary nature of private markets. Asset managers are likely to be extremely reluctant to share their data and other transactional information openly. Even if the data were anonymized and shared through secure APIs, the concerns over legal risks and competitive advantages would be substantial. These dynamics form a substantial barrier to the open integrations that would be necessary for creating frequent AI-driven valuation models.

Other systemic challenges are the stringent regulations inherent to the financial industry, especially regarding asset valuation practices. Many regulatory bodies require valuations to meet fair value measurement standards, primarily governed by accounting principles such as ASC 820 in the United States, and IFRS 13 internationally. These and similar standards often demand human judgment and oversight, calling into question if AI-generated intra-quarter marks based on comparable assets would satisfy such regulatory criteria.

Conclusion

As more institutional investors allocate capital to private markets, there's a greater demand for transparency and frequent reporting similar to what they're accustomed to in public markets. Moreover, regulators are also pushing for greater transparency in private markets. Finally, the private equity industry has occasionally shown a willingness to collaborate on data standardization and sharing. For example, in 2022, Carlyle and the California Public Employees’ Retirement System (CalPERS) collaborated to form the ESG Data Convergence Project, the private equity industry’s first-ever collaboration to standardize ESG metrics.

For now, the jury is still out on whether generative AI will truly be able to generate reliable private market intra-quarter marks using private market comps. However, if alternative investment firms can gain a competitive edge by doing so, the development and adoption of AI-driven valuation tools is likely to continue apace.

If you are interested in developing more in-depth knowledge of the potential of alternative investments, alternative please consider joining us at one of our upcoming Focus on Alts Series events: New York (October 23, 2024) and Los Angeles (November 14, 2024).

For financial advisors seeking a truly comprehensive mastery of alternative investments, Investments & Wealth Institute’s Certified Investment Management Analyst (CIMA) certification offers a world-class curriculum, covering a full spectrum of alternative assets, from private equity and private debt, to infrastructure, digital assets, and more. The CIMA program’s learning objectives are designed to sharpen expertise in this evolving asset class and provide advisors with the sophisticated knowledge and tools necessary to guide their clients effectively.

For further reading, you may be interested in Investments & Wealth Research (Issue 3, 2023) “Adapting to Growing Private Markets: A Playbook for Practice Success.”