AI - All The Work!

n this RAISE Summit talk, Ori Goshen—Co-founder & Co-CEO of AI21—reframes AI’s impact on knowledge work and unveils Maestro, AI21’s enterprise-grade agent for work. Building on Peter Drucker’s idea that modern economies are run by knowledge workers, Ori explains why traditional LLMs stumble on multi-step tasks (small error rates compound) and how Maestro blends the flexibility of prompting with the reliability of validated, step-wise execution—so organizations can automate mission-critical workflows with confidence.
‍

Key discussion points include:

• From tasks to outcomes: Work = billions of tasks; robust automation needs planning, execution, and per-step validation.
• Why agents fail: Single-step accuracy isn’t enough—multi-step chains amplify errors; static chains are rigid and slow to build.
• Meet Maestro: A model-agnostic agent platform where developers define instructions, tools, requirements, and compute budget.
• Trust & control: Validated outputs, confidence scores, and a transparent execution graph for auditability and governance.
• Planning at scale: Maestro generates multiple plans, runs parallel retrieval/model strategies, and selects optimal actions.
• Enterprise use cases: Examples include compliance & regulation analysis, customer support, and other knowledge workflows.
• Beyond hype to value: AI21’s thesis of super productivity—doing far more with the same teams—and real benchmarks that boost off-the-shelf model accuracy.
• What’s next: Moving from AI that “answers” to AI that gets work done safely, consistently, and at enterprise scale.

Interested in enterprise agents, validated workflows, and model-agnostic orchestration? Watch now—and like, comment, and subscribe for more sessions from the RAISE Summit.