Jyoti Phogat
Founder and AI Strategy Researcher | Trustworthy AI, Decision Intelligence and Agentic Systems
Ravenna, Ohio, United States
Actions
Jyoti Phogat is the founder and AI strategy researcher behind Lotus & Dorje and Student-LAD, an AI-supported platform designed to help students and institutions manage planning, decisions, documents, reminders, and workflows in secure, context-aware environments.
Her work focuses on translating AI strategy into governed and verifiable implementation. She is developing frameworks for strategic AI initiative selection, multi-level value measurement, agentic software delivery, continuous context evolution, and evidence-based verification. Her research examines how organizations can choose the appropriate level of AI autonomy while protecting sensitive information, maintaining human accountability, and connecting technical performance to measurable educational and operational outcomes.
Through the applied development and testing of Student-LAD, Jyoti identified the “false-success” problem: an AI system can communicate that an action was completed even though the authoritative application state remains unchanged. Her work turns this implementation challenge into practical guidance for education executives, project managers, technology leaders, developers, and AI governance teams.
Badges
Area of Expertise
Topics
The AI Said It Was Done—but Was It? Five Evidence Gates for Safer AI Use in EMS
Generative AI is rapidly entering emergency services through documentation assistance, training development, protocol retrieval, scheduling, quality improvement, public education, and administrative workflows. These tools can save time, but they can also produce confident answers that are incomplete, inaccurate, unsupported, or never reflected in the authoritative system.
Consider a seemingly simple failure: an AI assistant reports that a task, report, or record was updated, but the underlying system remains unchanged. In EMS, this gap between what an AI says and what actually happened can affect documentation, operational readiness, quality assurance, privacy, and potentially patient safety.
Designed primarily for EMTs and other frontline providers, this interactive session translates AI governance into a practical field-level method. Participants will examine realistic EMS scenarios and learn to determine which AI-supported tasks are appropriate, which require verification, and which must remain under qualified human control.
The session introduces five progressively stronger evidence gates:
Source Gate: Can the information be traced to an approved protocol, policy, record, or authoritative source?
Fact Gate: Have names, dates, times, clinical facts, and operational details been checked?
Action Gate: Did the requested action actually occur in the authoritative system?
Privacy Gate: Was protected, sensitive, or operational information handled appropriately?
Human Authority Gate: Was the output reviewed by the person legally and professionally responsible for the decision?
Through audience polling and short “Would You Trust This?” scenarios, attendees will evaluate AI-generated documentation, training content, protocol summaries, scheduling actions, and operational recommendations. The session will also distinguish low-risk uses—such as brainstorming public-education content—from higher-risk uses involving patient information, clinical decisions, medication guidance, or official records.
Participants will leave with a reusable one-page AI verification checklist that can be applied at the station, in the classroom, or within an EMS agency. The purpose is not to turn EMTs into AI engineers. It is to give EMS professionals a clear method for recognizing unsafe reliance, verifying useful output, protecting sensitive information, and keeping qualified humans accountable.
From Cloud-Only to Edge-First AI: Measuring What We Gain—and What We Risk
Cloud AI is powerful, but it is not always the right place to process every request. In manufacturing, healthcare, education, energy, and other distributed environments, organizations may need faster responses, stronger data locality, predictable costs, and continued operation when connectivity is limited.
This session presents a practical way to evaluate whether an AI workload should remain in the cloud, run on a local edge device, or use a hybrid architecture. Drawing on an ongoing, self-funded model-compression study, I will explain how quantization and teacher-guided recovery can be evaluated for more than accuracy alone. The discussion will connect model capability with memory use, latency, energy demand, hardware requirements, operational resilience, and total cost of ownership.
We will examine an AWS-supported hybrid pattern in which routine or sensitive inference runs locally, while cloud services support secure synchronization, monitoring, model updates, recovery, and escalation to more capable models. I will also discuss the limits of local AI, including device management, model drift, cybersecurity, update integrity, and the danger of allowing probabilistic models to control safety-critical operations.
Attendees will leave with a decision framework for comparing cloud-only, local-only, and edge-first hybrid deployments; a set of measurable technical and business criteria; and practical governance checkpoints for moving from an experiment to a responsible pilot. The session is educational and based on research in progress. It does not promote a product or assume that edge deployment is automatically cheaper, safer, or more sustainable.
From AI Strategy to Verified Action: A Leadership Framework for Trustworthy Educational AI
Educational institutions are moving from generative AI experimentation toward systems that can manage assignments, schedules, communications, documents, advising, and institutional workflows. Yet an AI assistant may confidently report that it completed an action even when the authoritative record was never updated.
This panel begins with a real failure identified during the development and testing of Student-LAD: an assistant reported that an overdue task had been completed, while the stored task remained unchanged. This gap between conversational confidence and operational truth exposes a critical leadership question: How can educational institutions verify that AI systems are reliable, secure, accountable, and connected to measurable outcomes?
Panelists will examine a practical leadership framework for evaluating AI initiatives, selecting the appropriate level of autonomy, and establishing evidence gates before AI reports success. The discussion will compare deterministic automation, AI copilots, single-agent systems, multi-agent solutions, and human-controlled workflows.
Participants will learn how Verified Completion Rate, False Success Rate, authoritative-state validation, privacy controls, and outcome-linked measurement can strengthen institutional AI governance. Examples will include student reminders, advising, course planning, document support, calendar scheduling, and administrative workflows.
Attendees will leave with a reusable decision checklist for determining whether AI should perform a task, how much autonomy it should receive, what evidence must prove completion, and what educational or operational outcome should justify the investment.
AWS DMV Community Day 2026 Sessionize Event
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top