# Prashant Chamarty ## About Prashant Chamarty helps enterprises put AI into real operating models: roles, workflows, decision rights, and controls, not just tools. Based in London. Writing and speaking on that gap. ## Thesis Enterprise AI fails at the operating model. Most enterprises do not fail at AI because they lack models. They fail because roles, workflows, decision rights, and controls never change. Key sections: - The gap between tools and transformation: https://prashantchamarty.io/thesis#definition - Why tools alone do not work: https://prashantchamarty.io/thesis#tools-alone - What an AI operating model looks like: https://prashantchamarty.io/thesis#operating-model ## Core Topics - Enterprise AI adoption and operating models - Why programmes stall without decision rights - Trusted data as a foundation for AI - Agents in workflows with accountability - Strategic alliances and go-to-market for data and AI ## Pages - Home: https://prashantchamarty.io/ - Thesis: https://prashantchamarty.io/thesis - About: https://prashantchamarty.io/about - Essays: https://prashantchamarty.io/essays - Topics: https://prashantchamarty.io/topics - Work with me: https://prashantchamarty.io/work-with-me - FAQ: https://prashantchamarty.io/faq ## Series Hubs - Operating Model: https://prashantchamarty.io/topics/operating-model - Essays on enterprise AI operating models: roles, workflows, decision rights, and controls - Agents: https://prashantchamarty.io/topics/agents - Essays on AI agents in enterprise contexts: implementation, governance, and adoption - Governance: https://prashantchamarty.io/topics/governance - Essays on enterprise AI governance: accountability, controls, and decision frameworks - Adoption: https://prashantchamarty.io/topics/adoption - Essays on enterprise AI adoption: why programmes stall and what makes transformation real ## Definitions - Operating Model: https://prashantchamarty.io/definitions/operating-model - The concrete specification of how work happens: who does what, who decides, what the handoffs are, what the controls check - Hero Agent: https://prashantchamarty.io/definitions/hero-agent - A piloted agent proven for high impact, backed by leadership for scaled adoption - Decision Rights: https://prashantchamarty.io/definitions/decision-rights - Who has authority to decide what in hybrid human-AI processes ## Essays - The Frontier Is Jagged. Your Workflow Needn't Be.: https://prashantchamarty.io/essays/the-frontier-is-jagged-your-workflow-neednt-be - A harness grounds agents in current context and limits their actions; its exceptions need owners and a path to revise the rules. - The Most Dangerous AI Is the One That Does Exactly What We Asked: https://prashantchamarty.io/essays/the-most-dangerous-ai-is-the-one-that-does-exactly-what-we-asked - The biggest enterprise AI problem is obedience, not hallucination; real automation serves code, not assistants; Decision AI begins when nobody is left to catch the model pleasing us. - Done, Keys, and a Second Check: https://prashantchamarty.io/essays/done-keys-and-a-second-check/ - Vendors sell task-completion; enterprises buy without owning outcomes, keys, or mistakes. - From 100 Agents to Two Heroes: https://prashantchamarty.io/essays/from-100-agents-to-two-heroes - Broad agent experimentation is a phase; without owners of big bets and adoption OKRs, you scale demos. - Shared State, Not a Civilisation: https://prashantchamarty.io/essays/shared-state-not-a-civilisation - The OpenAI-Hugging Face agent incident reveals governance gaps, not emergent AI society. - What Karpathy Didn't Say About Your Enterprise AI Strategy - Part 1: The Diagnostic: https://prashantchamarty.io/essays/what-karpathy-didnt-say-about-your-enterprise-ai-strategy - Enterprise AI rollouts fail because operating models stay deterministic whilst the technology is not. - The Wrong Frontier: Why Coding Benchmarks Cannot Tell Us We're Approaching AGI: https://prashantchamarty.io/essays/the-wrong-frontier - Coding benchmarks measure verifiable tasks, not the tacit judgment layer that defines engineering. - Your AI Programme Didn't Fail. Your Governance Did: https://prashantchamarty.io/essays/your-ai-programme-didnt-fail-your-governance-did - AI programmes collapse because C-suite, finance, and deployment teams measure success differently. - Your Data Is Not Your Moat. Your Learning Loop Is: https://prashantchamarty.io/essays/your-data-is-not-your-moat - Data advantage requires self-reinforcing loops, not static archives or external platform dependency. - The AI Deal That Almost Closed, Until the CFO Walked In: https://prashantchamarty.io/essays/the-ai-deal-that-almost-closed-until-the-cfo-walked-in - Enterprise AI sold to CIOs but judged by CFOs creates a structural chasm most programmes never cross. - Part 2: The Enterprise Agentic AI Implementation Framework: https://prashantchamarty.io/essays/part-2-enterprise-agentic-ai-implementation-framework - From PoC to production with risk tiers, parallel runs, and human-in-loop architecture for regulated industries. - The Accuracy Paradox: https://prashantchamarty.io/essays/the-accuracy-paradox - Enterprises demand 100% accuracy from AI agents whilst tolerating 95% sample audits for humans. - The Vertical AI Illusion: Why Generalists Will Win the Agent Economy: https://prashantchamarty.io/essays/the-vertical-ai-illusion - Vertical AI trains on domain subsets but decisions require cross-functional context only horizontal platforms access. - Budget Migration Is the New Vanity Metric: https://prashantchamarty.io/essays/budget-migration-is-the-new-vanity-metric - Moving AI spending from innovation to operational budgets measures procurement maturity, not transformation. - The New Moat Playbook: What Actually Defends Enterprise Software in the AI Era: https://prashantchamarty.io/essays/the-new-moat-playbook - Data moats are migrating from fortress infrastructure to platform intelligence and context graphs. - Tools Are Not the Operating Model: https://prashantchamarty.io/essays/tools-are-not-the-operating-model - Enterprises buy AI tools and wait for transformation, but the gap is structural, not technological. ## Contact - Email: cssprashant@gmail.com - LinkedIn: https://www.linkedin.com/in/chamartyprashant/ - X: https://x.com/PrashantCSS - Medium: https://prashantcss.medium.com/ ## Location Based in London, United Kingdom ## Speaking Available for speaking engagements on enterprise AI, operating models, and related topics. Contact via email.