RelationalAI

Optimization in financial services has a reach problem. What's missing is a shared layer that lets people express decisions and constraints without going back to rebuild models each time the question changes. RelationalAI provides a new way: an ontology over the financial data with every analysis running on top of a shared substrate. Each reasoning step feeds the next: → Rules flag concentration risk to shape constraints → Graph analysis surfaces connector stocks with the largest exposure to weight the objective → Optimization helps us trace the efficient frontier and surface key trade-offs → Stress test scenarios show which portfolios hold up in a crisis The result worth flagging: the "safe" portfolio under normal conditions flipped to risky under a correlation spike. The middle of the efficient frontier held up across scenarios. When the formulation lives next to the data and reasoners compose on a shared model, the gap between it and the business question closes. https://lnkd.in/gfYba64s

A frontier model is like a freshly minted graduate: brilliant on paper, but on day one it doesn't know your fiscal calendar, your product taxonomy, your customers, or how decisions actually get made. That gap is what post-training closes. Through fine-tuning, distillation, and reinforcement learning on your private data and context, a smaller open-weight model can become the professional who already knows the business and handles the majority of enterprise use cases at a fraction of frontier costs. Our CEO Molham Aref makes the case for the multipliers that make enterprise-scale agent economics actually work.

When are we going to have agents running our business? We now have coding agents that produce code, co-work agents that produce documents and spreadsheets, and design agents that generate wireframes and visualizations. What do we need to have agents that do revenue management, merchandising, supply chain planning, and other important business functions? I wrote a piece on what I think actually brings those agents to life: context, tools, and post training. Would love your thoughts. https://lnkd.in/gUwACVt8

Vinay K. Chaudhri moderates the “Computational Law as a Blueprint for Ai Reasoning” panel at Stanford Law School’s Codex #FutureLaw conference. Joined by Molham Aref of RelationalAI, Patrick McLaughlin with the Hoover Institution, Stanford University, Abhijeet Mohapatra of Symbium, and Marzieh Nabi, PhD with Codex. Observations in comments.

📣 Two 2026 SIGMOD Research Highlight Awards on languages and query optimization:  📄 Rel: A Programming Language for Relational Data 👉 https://lnkd.in/gV9fKq_c (Molham Aref, Paolo Guagliardo, George Kastrinis, Leonid Libkin, @Victor Marsault, Wim Martens, Mary McGrath, Filip Murlak, Nathaniel Nystrom, Liat Peterfreund, Allison R., Cristina Sirangelo, Domagoj Vrgoc, David Zhao & Abdul Zreika)                                                                                                            📄 The Key to Effective UDF Optimization: Before Inlining, First Perform Outlining 👉 https://lnkd.in/gxkB7rhY (Sam A., Yuchen Liu, Todd Mowry, Jignesh Patel & Andrew Pavlo) 👉 2026 SIGMOD Research Highlights: https://lnkd.in/gqbaSRe3 #SIGMOD2026 #ACM #databaseresearch

As far as industries go, few are more intrinsically primed to lead the shift from reporting to operationalizing data into revenue-driving decisions using Decision Intelligence than Telco. (Or Techco, depending on which side of the chasm you sit). Closing the loop from "what happened" to "here's what you should do next" however, is no easy task; here's a real-life demo by our own Ankit Tandon on how RelationalAI combines knowledge graphs, AI reasoning, and natural language to get: > root cause analysis across complex systems, > churn prediction for high-value customers, > prescriptive optimization for all operations. All powered natively, inside Snowflake—no data movement, fully governed, and explainable. For more info on our technology, check out our blog 👇

In my recent webinar hosted by Snowflake, I teamed up with Fawad A. Qureshi to showcase how data teams can move from semantic views and analytics to decision intelligence by augmenting the semantic layer with business ontologies and advanced reasoning from RelationalAI — all natively inside Snowflake. If you missed it, you can watch the full session on-demand here: https://lnkd.in/g8jrQdna

Most companies have, by now, invested in an AI data cloud. Now they're looking at ways to turn data into real and timely decisions faster, and with less fragmentation. Using BI tools answers questions like “What happened?”, but rarely “What will happen?” or “What should we do?”—so the gap to decisions remains. That’s the loop we focused on closing in our recent session with Snowflake, "From Data Chaos to Unifying Data, Business Logic and Decisions": how to harvest decision intelligence step by step, from data to decision, all inside the AI Data Cloud. Our own Ankit Tandon and his co-presenter, Fawad Qureshi, walked attendees through: · how semantic layers + ontologies ground business context · how reasoning (graph, predictive, rules, prescriptive) works together · what it looks like to model and optimize decisions directly in Snowflake · how Snowflake Intelligence + RelationalAI make this accessible to all users within the organization without requiring specialized expertise Our demo showed how RelationalAI connects knowledge graphs, business rules and reasoning to enable decision-aware agents across real-world use cases in supply chain, marketing and operations. If you missed it, you can watch the full session on-demand here: 👉 https://lnkd.in/gfmudxeY

Everyone's racing to build AI data centers, but the grid can't keep up. Watch an agent powered by RelationalAI forecast demand → check feasibility → analyze criticality → optimize data center approvals and substation upgrades per investment scenario. It's all grounded in the ontology and powered by structured reasoning, each step building on the last. https://lnkd.in/gsTBEg4k

Multiple reasoning approaches surface risks no single analysis catches. Supply chain data in Snowflake → RelationalAI ontology → chained reasoning: graph analysis found bottleneck sites, rules flagged high-risk suppliers, optimization routed inventory to minimize cost to meet demand. Then a what-if disabled a "safe" supplier – only to reveal a hidden point of failure. These questions require more than SQL and LLMs. The ontology made this fast – and kept the agent grounded in real business logic and structured reasoning instead of... creative assumptions. https://lnkd.in/gdbKbG7P