Abhilash Thumma, Enterprise Architect, and Ashwin Mishra, Head of AI Services (FSI, P&E), presented “Finsights: Financial Insights and Analysis Using AI Agents” at Google Cloud Next 2026. We worked directly with Abhilash and Ashwin when Google Cloud contracted us and our data to build FinSights.
Google Cloud chose us for building FinSights because we are the ONLY source of accurate financial data and analytics that verifiably produce alpha in the stock market.
After the showcase, Ashwin noted
“Data without actions is well, just data… Everything without action is a distraction and we could not have done it without this amazing team. Special thanks for the design partnership, David Trainer and team – you guys are amazing.”
Their presentation covered:
- how AI agents are moving beyond basic chat to perform complex, multi-step financial analysis,
- the unique challenges of financial data, and
- how to build agents that teams can actually trust.
Below, we summarize some of the key takeaways from the session.
Financial Analysis: Moving Beyond LLMs to Small Language Models
By now, most everyone knows how difficult it is to get machines to analyze financial filings and produce alpha-generating measures of earnings. If it were easy, we’d see lots of investors doing it. We’re not seeing that, but we are seeing the industry shift away from general large language models (LLMs) to specialized purpose-built tools, e.g. small language models. As Ashwin noted, “general purpose LLMs are great at writing emails…but when it comes to very specific financial time series information, they could find that a little bit challenging.” FinSights is built on superior data from our small language model so it can deliver alpha-generating insights. Details in the replay here.
Figure 1: The Shift to Specialized Purpose-Built Tools
Source: Finsights: Financial Insights and Analysis Using AI Agents Replay
FinSights Drives Real Intelligence
Because FinSights is built on better data, it is not just a tool, but rather a “digital extension of your investment team”, as Ashwin noted during the session. It can accurately answer questions, deliver actionable insights, and empower data-driven decision making. Watch this section of the replay here.
Figure 2: FinSights Drive Real Intelligence
Source: Finsights: Financial Insights and Analysis Using AI Agents Replay
Integrity Matters
FinSights drives real intelligence, not just because it’s powered with superior data, but because it is built to monitor and observe the integrity of its own answers. Abhilash put it best when he said, “sounding confident isn’t just enough, the underlying data must be flawless.”
Throughout the development process, we provided Google with human-verified reference prompts, or the exact correct answer to specific questions, aka the “Golden Prompt.” We also showed them how to find the answers in the data. With one version of the truth, the FinSights agent can validate its answers against those golden prompts to ensure accuracy and integrity. Watch this section of the replay here.
Figure 3: Quality Control Improves the Integrity of the Agent
Source: Finsights: Financial Insights and Analysis Using AI Agents Replay
Accurate Data Drives Alpha
Google Cloud built FinSights to showcase the art of the possible when their AI is powered by accurate data. At 29:55 in the replay, you can see our ROIC data featured in the presentation.
Google Cloud chose us for building FinSights because we are the ONLY source of accurate financial data and analytics that verifiably produce alpha in the stock market. Learn more about FinSights here.
Figure 4: New Constructs’ ROIC Data at Google Cloud Next 2026
Source: Finsights: Financial Insights and Analysis Using AI Agents Replay
This article was originally published on May 1, 2026.
Disclosure: David Trainer and Kyle Guske II receive no compensation to write about any specific stock, sector, style, or theme.
Questions on this report or others? Join our online community and connect with us directly.