Savart CEO on how AI powers AD ASTRA fund for smarter long-term investing

Sankarsh Chanda, Founder and CEO of global asset management firm Savart, says the company's AD ASTRA Fund -- driven by its APART AI -- leverages a distinctive three-engine framework to build and manage high-conviction, long-term portfolios. In a conversation with Business Today, Chanda explains that the process integrates deep financial analysis (Quant), qualitative assessment (Iris) and strategic decision-making (Synapse).

He notes that APART AI responds in real-time and demonstrated its effectiveness by safeguarding capital during the Covid-19 market crash. Unlike conventional funds, Chanda stresses, AD ASTRA combines data-driven discipline and adaptability -- positioning it as a compelling choice for long-term investors. Here's an edited excerpt from the interview with Savart CEO:

* How does Ad Astra's APART AI model select and weigh stocks for long-term investment?

Sankarsh Chanda: The AD ASTRA Fund, powered by APART AI, uses a three-engine system to build and manage long-term, high-conviction investments. The Quant Engine handles quantitative analysis, scanning over 2,000 financial parameters across P&L, balance sheets and cash flows to identify fundamentally strong companies. The Iris Engine focuses on qualitative factors like corporate governance, promoter ethics and sustainability by analysing over 4,000 sources including reports, calls and social media. The Synapse Engine integrates insights from both to make portfolio decisions -- stock selection, weighting and rebalancing -- based on real-time conviction and risk. This structured yet flexible approach allows the fund to stay aligned with its long-term growth philosophy while adapting quickly to market changes.

* What safeguards are in place to prevent overfitting or bias in the AI's decision-making process?

Sankarsh Chanda: Bias is dangerous and APART AI is built to lower it at all costs. Unlike traditional investing where decisions are often influenced by human bias or historical heuristics, APART AI does not form opinions before conducting research. It approaches every company and opportunity with a clean slate. It is sector-agnostic and non-conforming; it doesn’t follow rigid classifications or standard views of what defines a good company or industry.

APART AI avoids conventional valuation traps. It doesn’t pre-assume what a "good" PE ratio is, eliminating common biases that even seasoned fund managers fall into. Our system separates outcomes from assumptions. It avoids attaching reasons to results unless supported by data, reducing the risk of confirmation bias or narrative fallacies. This rational, evidence-based framework ensures that the portfolio is not driven by trends, emotions, or flawed human intuition, making APART an ideal long-term steward of investor capital.

* How does AD ASTRA plan to tackle competition?

Sankarsh Chanda: While AI isn't perfect, the APART AI system gives AD ASTRA a distinct edge over traditional fund managers by addressing their most common weaknesses. APART, by contrast, operates with no emotional bias, no cognitive fatigue, and a far broader analytical bandwidth. APART stays continuously engaged, scanning and adjusting without the psychological urge to "exit and wait," giving it better odds of capturing critical return bursts. It doesn't rely on intuition, assumptions, or "gut feeling." It processes thousands of data points 24x7 across financial and qualitative dimensions and gives it an information advantage unmatched by human teams.

In short, AD ASTRA doesn't compete by imitating others; it redefines the approach entirely, using technology-led consistency, discipline, and scale to outperform.

* Can you provide a few cases of how the AI has responded to past market downturns or anomalies?

Sankarsh Chanda: One of the most notable instances where our AI system — APART — demonstrated its effectiveness was during the Covid market downturn. In early 2020, while we were managing approximately Rs 100 crore in assets, the system issued a strong alert recommending a shift to liquid assets. At the time, there was no widespread panic in the Indian market, and even we didn't fully understand the rationale behind the alert. However, trusting the system's data-driven insights, we moved nearly 65 per cent of the portfolio into liquid instruments by February 2020.

Just a month later, Covid hit India and global markets began to plummet. What seemed like a cautious move turned out to be a critical decision that preserved investor capital during one of the most volatile periods in recent history.

It's important to note that APART had been tracking the earliest global signals of the pandemic as far back as 2019, including developments outside India. Based on its continuous analysis of global risk factors and sentiment shifts, the system identified the potential for a major negative market event and acted well in advance.
This proactive reallocation helped protect close to Rs 100 crore of investor wealth. Post the market correction, we re-entered strategically enabling us not only to preserve capital but also to generate strong returns in the recovery phase. 

*  How is it different and why should investors select AD ASTRA?

Sankarsh Chanda: AD ASTRA is fundamentally different from conventional quant or actively managed funds. While most quant funds focus purely on technical indicators and mathematical models, APART goes further and conducts in-depth fundamental analysis (balance sheet, PCL, cash flows, etc.) and high-scale qualitative research (corporate governance, ethics, sustainability, promoter behaviour).

APART AI analyses over 2,000 financial metrics and taps into 4,000-plus qualitative sources in real time. AD ASTRA doesn't wait for ideal macro conditions, instead, it combines top-down market insights with bottom-up company fundamentals to capture opportunities, even in challenging environments. It's a fund built for long-term investors.