State of the Art in Agentic AI
summary
Claims a working multi-agent consumer assistant called Kriti that launched last week and supports English +11 languages, is on Play Store iOS and web, and uses model context protocols (MCPS) plus an agent-to-agent protocol to coordinate agents for ordering, payments and tracking. Architecture details: multimodal input/output, orchestrator that decomposes tasks into subtasks, verifier to confirm outcomes. Contrast: browser agents take 5–15 minutes and often fail on transactions; Kriti is presented as functional for real transactions via MCPs like Amazon and PayPal.
AI-written summary · about
chapters (6)
- 0:00Introduction to agentic AI
- 1:12Levels of AI intelligence
- 2:16Defining AI agent types
- 4:44Multi-agent assistant demo
- 7:01Building true AI assistants
- 10:10Architecture of agentic systems
related talks
01h 01mKeynote Panel: Future of AI agents
28mThe Put That Sh*t On Agent: Building Multi-Agent AI Systems That Actually Talk to Each Other*
40mAgentic AI: The Next Evolution of Automation
29mAgentic AI: What It Is and Why It Matters
40mAI Agents in Practice
26mBuilding Agentic Workflows: Patterns for Orchestrating Intelligent Systems.
27mMulti-Agent AI: Transforming Business Operations and Value Creation
16mAgentic AI and What's Coming for Consumer Tech