关于凯洛
About Kailoo
一支墨尔本的 AI 工程团队。下面是我们的发展历程、AI 走到今天的几个节点，以及到目前为止做成的事。
An AI engineering team in Melbourne. Below: our history, the milestones that brought AI to where it is, and what we have delivered so far.
业务一览
What we do
六条业务线，各有专业名称，每条都对应可验收的交付物。
Six lines of work, each with a defined deliverable that can be accepted.
-
数字产品工程
Digital product engineering
网站前端设计与全栈开发
Front-end design and full-stack development
了解方案 See the approach -
智能体与知识系统
Agents and knowledge systems
RAG 智能体：客服、售后与内部知识问答
RAG agents: customer service, after-sales and internal knowledge Q&A
了解方案 See the approach -
增长与电商运营
Growth and e-commerce operations
线上战略规划与执行：品牌、主图、Listing、转化与选品
Online strategy, planned and executed: brand, hero images, listings, conversion and product selection
了解方案 See the approach -
内容与 AIGC 生产
Content and AIGC production
AI 自媒体运营与定制化 AIGC 内容
AI-run social accounts and custom AIGC content
了解方案 See the approach -
AI 智能运维
AI-driven operations
邮件与邮箱、社媒账号、搜索与 AI 可见性的整理与运营
Email and mailboxes, social accounts, search and AI visibility, organised and run
了解方案 See the approach -
语音与数字人
Voice and digital humans
AI 语音电话与 AI 数字人直播
AI voice calls and AI digital-human livestreaming
了解方案 See the approach
发展历程
Our history
-
2026.07
gig-radar 开始无人值守运行：多源采集、模型评分、分渠道起草、回复扫描。
gig-radar starts running unattended: multi-source sourcing, model scoring, drafting per channel, reply scanning.
-
2026.08
kailoo.ai 上线。
kailoo.ai goes live.
-
2026.09
福木源紫檀博物馆官网上线；AI 科普视频生产线跑通整集；grounded-chat 通过 223 项自动化测试；本站改版并接入站内问答。
The Fumuyuan museum site goes live; the AI explainer video line runs a full episode; grounded-chat passes 223 automated tests; this site is redesigned with the on-site assistant.
AI 发展史记
A short history of AI
我们所处的位置，放在七十年的时间线上看。
Where we stand, on a seventy-year timeline.
-
1956
达特茅斯会议提出「人工智能」这一名称。
The Dartmouth workshop gives the field its name: artificial intelligence.
-
2012
深度学习在图像识别上取得突破，神经网络重新成为主流。
Deep learning breaks through in image recognition; neural networks return to the centre.
-
2017
Transformer 架构提出，成为此后所有大语言模型的基础。
The Transformer architecture is proposed; every large language model since builds on it.
-
2022
ChatGPT 发布，对话式 AI 进入公众视野。
ChatGPT is released; conversational AI reaches the public.
-
2023
多模态模型出现，AI 开始同时处理文字、图像与语音。
Multimodal models arrive; AI handles text, images and speech together.
-
2024
智能体与视频生成走向可用，AI 从「回答」走向「做事」。
Agents and video generation become usable; AI moves from answering to doing.
-
2026
推理模型与智能体进入企业的日常工作流程。我们在这一年做了上面这些事。
Reasoning models and agents enter everyday business workflows. This is the year we did the things above.
到目前为止
So far
- 4个系统上线运行：博物馆官网与三套自研systems live: the museum site and three in-house
- 6条业务线，一支团队统一承接lines of work under one team
- 284项自动化测试通过（223 + 61）automated tests passing (223 + 61)
- 809条线索经模型评分leads scored by the model
- 21镜，一集科普视频整链跑通shots in one explainer episode, full chain
- 16个路由，博物馆官网上线routes on the museum site, live