newsfilter.io
Interview, Fireside Chat

Agents, Lawyers, and LLMs

  • O1 and future models are not expected to produce complete S1 outputs in a single "one-shot" attempt, necessitating reliance on iterative processes, interactive feedback loops, and "shoulder taps" for complex legal tasks.
  • The company plans to expand vertically into tax, financial diligence, HR, and compliance roles over the next two to three years by leveraging workflow contexts from existing legal projects to attract non-legal professionals.
  • Adoption of technology is projected to accelerate among law firms to signal innovation and improve efficiency, driven by competitive dynamics and client pressure, though widespread ROI remains constrained by human and software bottlenecks regarding trust and workflow integration.
  • Market pricing is expected to shift toward outcome-based models over time, but current enterprise buyers face challenges in valuing experimental outcome-based work until evaluation metrics evolve to focus on completion percentages (e.g., 85-90%) rather than binary accuracy.
  • Infrastructure strategy involves continuing reliance on major model providers like OpenAI and Azure for foundation models and infrastructure to avoid cost and expertise constraints associated with building proprietary models.
  • A strict "no training" policy for customer data will be maintained by default to preserve enterprise trust, while latency is anticipated to decrease as AI output quality and human-like nature improve.
  • The market is projected to see innovation in AI-native UX and dynamic interfaces in 2025, moving beyond text-based chat toward specialized, workflow-specific experiences, though most users are not expected to achieve full efficiency or reach an "AGI heaven" scenario within the next two to three years.
  • Enterprise buyers are expected to require proactive agents capable of coordinating collaboration by pinging different personnel at each firm, and future billing and staffing models will continue to evolve with AI adoption despite specific changes remaining uncertain.
  • Legal expertise will remain embedded across all company functions, including sales, product, and AI engineering, while most law firm customers are expected to learn of technological shifts via social channels like LinkedIn rather than Silicon Valley news cycles.