Hi, I'm Sandesh Sathyanarayana.
I connect the dots, bringing humans front & center, in the world of AI.

My POV on AI

As I concluded my current VC Fellowship with Metvy, recently, I often ask myself, what do I look for in the next wave of Frontier AI. While I am not an investor, my 17 years in the digital trenches -from Google to CleanTech - have taught me to look past the "Generative" hype and identify the Operating Moats.I am betting my professional reputation on the transition from AI as a "tool" to AI as a B2B2C Platform. These are the five theses I believe will define the next decade of the power law, and they are exactly where I am focusing my strategic advisory today.

1. Commerce: The Decision OS
Traditional CRM data is often "frozen"—static logs that fail to capture current intent. My thesis is a shift toward a Decision OS where Agentic RAG (Retrieval-Augmented Generation) activates this data. By triangulating real-time signals, we move from disruptive, auction-based auctions to empathetic "customer nudges" that prioritize long-term value.2. Sustainability: The Agri-Carbon Ledger
Carbon markets are currently hindered by a "manual chasm." While I am not the one writing the CV algorithms, I see the strategic path to automating audits. By using Edge Vision AI for biomass assessment and connecting it via API to certificate trading platforms, we can give small-scale farmers biological sovereignty and tradable assets in a high-velocity green economy.3. Energy: Solar Governance & Neutrality
The bottleneck in green energy isn't the engineering; it’s the distrust in multi-stakeholder decisions.The Architecture: A brand-agnostic platform powered by Google that ingests multiple vendor proposals.The Result: Using Multimodal LLMs as a neutral Single Source of Truth, allowing community residents to ask site-specific questions and reach consensus transparently.4. Media: Dynamic Stream Divergence
Linear viewership is colliding with a global attention deficit. By utilizing Vertex AI Video Intelligence, we can architect "sub-streams" that follow a viewer's contextual curiosity via a PiP interface. For brands, this transforms advertising from an interruption into a Trust-Score powered nudge, grounded in the immediate moment of interest.5. Performance: The Intelligent Edge in Sports
My approach to HaaS (Hardware-as-a-Service) for sports focuses on operational durability and unit economics. Instead of fragile sensors on moving gear (bats/balls), the strategic play is in Stationary Intelligence - stumps, nets, and pillars. This minimizes technical failure points while providing the high-fidelity data needed for real-time ML coaching and strategic benchmarking.The Orchestration Layer: MCP & A2A
To make these use cases scalable, we must look at the delivery layer. I’m particularly interested in how the Model Context Protocol (MCP) and Agent-to-Agent (A2A) platforms will allow specialized agents to collaborate autonomously. This isn't just about "using an API"; it’s about architecting an ecosystem that is flexible enough to evolve as the underlying tech stack does.Conclusion: The Bridge
My role isn't to be the CTO - I'm not qualified for that - it’s to be the Strategic bridge between Tech and Business, one who understands the potential of the Lakehouse, the Edge, and the Agent, and knows how to align them with a GTM strategy.The most helpful AI is the one that stays invisible while making the human at the center more powerful. As a Strategic Digital Thought Leader, I am looking for the dots that - when connected - don't just create a product, but define a new category.