Nasdaq has rapidly expanded the use of AI agents over the past 18 months, deploying them across market surveillance, compliance, and market microstructure analysis while keeping humans in the final approval loop. Pranav Ramesh, head of options research at Nasdaq and co-founder of AI startup Leadpoet, said the most notable shift has been in trust, as earlier AI systems often “hallucinated” too frequently for enterprise workflows. Ramesh expects crypto trading platforms to become early leaders in implementing AI agents for both internal operations and retail-facing tools, including position analysis, trade suggestions, and execution support, though full autonomy is not yet the model in practice.
At Nasdaq, AI agents automate low-value, high-volume tasks, particularly in anti-money laundering workflows through Nasdaq Verafin’s Agentic AI Workforce. The exchange’s Dynamic M-ELO order type, launched in 2023 and approved by the SEC, uses an AI model analyzing over 140 factors to adjust to real-time market conditions. Ramesh believes these institutional experiences offer a blueprint for crypto platforms, where machines can handle analysis and workflow, but humans remain the final checkpoint, ensuring oversight and reducing operational risk.
Ramesh also spoke candidly about labor impacts, noting that AI agents are already displacing lower-level roles. Positions in software, customer service, and analysis are being redefined or eliminated as companies deploy faster, more reliable AI. Recent examples include Crypto.com cutting 12% of its workforce, Messari restructuring under an “AI-first” model, and Block laying off 40% of staff citing improved AI systems. This trend underscores the transformative effect of AI on operational efficiency and workforce structure in financial and crypto sectors alike.
The labor shifts observed at Nasdaq and other firms informed Ramesh’s founding of Leadpoet, an AI-powered lead qualification startup. Leadpoet translates web signals and company context into “decision-ready” recommendations, focusing on precision over volume. The company uses Bittensor, a decentralized blockchain-powered AI network, enabling participants to contribute models and compute while earning rewards. Leadpoet has also joined NVIDIA’s Inception program, receiving technical support and ecosystem access, and achieved a $1 million annualized run rate in its first quarter after launch, signaling strong early adoption.
Ramesh sees the trajectory of AI adoption in crypto accelerating compared with other financial sectors. Autonomous agents are expected to handle significant operational tasks faster, while still requiring human oversight for final decision-making. The combination of institutional experience, decentralized AI frameworks, and retail trading demands positions crypto platforms to lead the integration of AI agents, reshaping operational workflows, improving efficiency, and redefining the interaction between humans and machines in trading and compliance environments.
