---
title: "Unique Friend | Future-Form’s Yang Jinsong: Agents Are the Hands and Feet of LLM Thought"
author: "Unique Research"
sourcePublication: "Unique Research (FFCAP English library)"
originalPublishedAt: "2024-01-19T16:30:00+08:00"
canonical: "https://ffcap.cn/en/research/20240119a06rij00"
source: "https://ffcap.cn/en/research/20240119a06rij00"
language: "en"
cnSourceUrl: "https://view.inews.qq.com/a/20240119A06RIJ00"
---

# Unique Friend | Future-Form’s Yang Jinsong: Agents Are the Hands and Feet of LLM Thought

_Unique Research · Insight / Unique Friend Q&A · January 19, 2024 · Guest: Yang Jinsong, founder & CEO, Future-Form / AutoAgents.ai (未来式智能)_

Inventory type remains **insight** (Unique Friend series — not partner welcome).

## Guest

**Yang Jinsong** — founder & CEO, Future-Form (AutoAgents.ai / 未来式智能). Former Feishu AI & Search product lead at ByteDance; previously led AWS aPaaS (~$200M annual revenue); SAP senior architect; Accenture consulting manager — clients include Sinopec, CNOOC, State Grid, Mengniu, New Hope, Xiaomi, Lenovo, FAW, BAIC, CRRC, TCL, Gree and peers.

## Q&A

**1\. Product and core scenes?**  
AutoAgents.ai’s flagship is the enterprise Agent app builder **Lingda (灵搭)** — an LLM app tool for business users. It aims to overturn rigid workflows, help firms build intelligent generic business flows, and create a human–machine co-creation paradigm — toward 10× productivity for 1B knowledge workers worldwide. As an enterprise tech builder, AutoAgents.ai serves government, power/energy, finance, and internet leaders with Agents that reshape business flows. Live scenes: sales assist, itinerary booking, contract review, customer ops, vertical knowledge-base Agents.

**2\. Opportunities vs. past waves?**  
When Agents are mainstream and best practices are highly SOPized, firms that need top-tier domain know-how may simply buy it for an extra fee. Language-model-centric apps will also deliver more scenes and better experience, overturning classical software’s accumulated debt.

**3\. Competing with giants?**  
Many AI startups sit on foundation-model tech; if a feature or scene can be covered after a few model-vendor iterations, the risk is high. AutoAgents.ai stays alert: product and service start from customer efficiency and experience, and deposit vertical-scene depth — so whatever the underlying model does, the response stays tied to customer pain.

**4\. Human–machine collaboration?**  
Agents build a “human–machine hybrid super pipeline.” Hybrid intelligence means people and machines finish a complex task together via workflows: hard work stays human; automated subtasks go to optimized Agents. Orchestrate hundreds of subtasks, SOPs, reviews, and QA into a mixed-intelligence line — like Ford’s assembly line (presses plus people tightening bolts machines could not yet do). Knowledge work becomes a knowledge-era assembly line of human + Agent collaboration — not dogmatic full automation.

_1 CN image; attach on ingest. Signup CTA in CN omitted._

— END —

---

Editorial English edition published on UniqueCapital: https://ffcap.cn/en/research/20240119a06rij00
On-site reading page: https://ffcap.cn/en/research/20240119a06rij00